14415_fritzsche_2874_r

Digital Frequency Analysis of Valve Sound Phenomenain Patients after Prosthetic Valve Surgery: Its Capabilityas a True Home Monitoring of Valve FunctionDirk Fritzsche1, Thomas Eitz1, Kazutomo Minami1, Delewar Reber3, Axel Laczkovics3, UweMehlhorn4, Dieter Horstkotte2, Reiner Körfer1 Clinics for 1Thoracic and Cardiovascular Surgery and 2Cardiology, Heart and Diabetes Center North Rhine-Westphalia, BadOeynhausen, 3Clinic for Thoracic and Cardiovascular Surgery, University Bochum, 4Clinic for Cardiovascular Surgery,University Köln, Germany Background and aim of the study: Depending on the
ed and examined meticulously.
individual risk profile of a patient, disturbances of
Results: Fourteen patients were found to have suspi-
the functional integrity of mechanical heart valve
cious signals. In 13 patients, valve dysfunction was
prostheses occur in up to 2.5% of patients each year.
confirmed by fluoroscopy, but in four cases neither
The early phase of prosthetic dysfunction (due to
transthoracic nor transesophageal echocardiography
thrombus formation, tissue ingrowth or endocardi-
detected abnormal occluder motion or ‘musses’ adja-
tis) usually remains undiagnosed, as patients do not
cent to the prosthesis. Normal valve sounds returned
present with symptoms in this situation, and imag-
in four patients who underwent thrombolytic thera-
ing techniques (echocardiography, fluoroscopy)
py. All patients regularly recorded and passed on
demonstrate normal occluder motion. The delay
their signals. Surveys revealed high acceptance and
between the onset of prosthetic valve dysfunction
easy handling of the Thrombocheck device.
and its clinical manifestation may result in complica-
Conclusion: Home monitoring of sound pressure
tions (e.g. thromboembolism) or extended therapeu-
measurements of prosthetic valves by digital fre-
tic options (e.g. reoperation rather than more
quency analysis via a Fast Fourier transformation
intensive anticoagulation).
may detect even very mild alterations of prosthetic
Methods: A total of 291 patients with mechanical
valve function. The next evolution of control sys-
heart valves was allocated to four different sub-
tems, allowing for registration of flow, frequency
groups, and each measured their valve sounds regu-
spectrum and electrocardiography, opens potential
larly with the ‘ThromboCheck’ device. Depending
applications for Internet-based, remote monitoring
on the subgroup, the signals were compared with
of cardiac patients.
different reference signals. Patients in whom a suspi-
cious signal was detected were immediately contact-
The Journal of Heart Valve Disease 2005;14:657-663 Despite major advances in the treatment of heart probably develops over a period of several days to valve lesions, there remains a complication rate of weeks. Prosthetic valve dysfunction is usually discov- between 0.8% and 6% per year for those patients who ered only after manifestation of serious fatal events, have undergone valve replacement surgery (1-4).
such as thromboembolic complications or heart failure.
Considering that approximately 100,000 heart valve In order to reduce the risk of undiagnosed valve dys- operations are performed worldwide each year, the function between two medical consultations, a system number of serious complications related to thrombo- was developed which allows recording of the acoustic sis, tissue ingrowth, leaflet tears, calcification, leakage phenomena with minimal effort at any place world- and prosthetic valve endocarditis still numbers several wide, and in a manner which is economically accept- A prosthetic, non-structural valve dysfunction which As minimal changes in prosthetic valve function is consistent with disturbed motion of the leaflets were reliably detected by frequency spectra analysis caused by thrombus formation or tissue ingrowth (5-8), this technique has the potential to diagnose dys-functions long before hemodynamic deterioration ormanifestation of systemic complications occur.
Initially, preliminary experience was gained with PD Dr. med. habil. Dirk Fritzsche, Abteilung für Thorax- undKardiovaskularchirurgie, Herz- und Diabeteszentrum Nordrhein- recording the frequency spectra of the click phenome- Westfalen, Georgstr. 11, D-32545 Bad Oeynhausen, Germany na of prosthetic valves, their analog/digital processing 658 Home monitoring of prosthetic valves and comparison with a reference file. The encouraging are shown in the display of the device with a short initial results obtained led to the introduction of a first message (‘ok’, error > ‘repeat measurement’, 2nd error hand-held device (the ‘ThromboCheck’) for home monitoring of the functional state of prosthetic heartvalves.
Subgroup 3
Five devices were utilized in an application trial in Clinical material and methods
cardiological centers in Israel. In this trial, the patient’sfrequency spectrum (valve models, all bileaflet, were This study comprised three aspects: simulation stud- SJM, CarboMedics, Sulzer Carbomedics (Austin, TX, ies, animal experiments, and some clinical investiga- USA) and Sorin Bicarbon (Sorin Biomedica, Saluggia, tions, the results of which have been published Italy) was compared with the 90% percentile of an previously (5-7). Herein are presented the preliminary average frequency spectrum obtained from 50,000 pre- results of the first multicenter application of the clinical measurements of the frequency and amplitude ‘ThromboCheck’, and the first single-center prospec- distribution of click sounds of different mechanical tive, double-blind study. After appropriate training for prosthetic valves. At present, 71 patients are involved use of the Thrombocheck, four different sub groups of patients were evaluated with the device.
Subgroup 4
Subgroup 1 (double-blind, prospective)
Another 102 patients (68 males, 34 females; 87 with Eighty-eight patients (51 males, 37 females; mean AVR, 15 with MVR; sinus rhythm in 83%) provided age 52.6 years) underwent aortic valve replacement themselves with a ThromboCheck device late after the (AVR) (n = 51), mitral valve replacement (MVR) (n = implantation of a mechanical valve prosthesis (mean 30), pulmonary valve replacement (PVR) (n = 1), MVR 4.2 ± 1.7 years), having been informed of the device’s + tricuspid valve replacement (TVR) (n = 2), AVR + existence via the Internet or self-help groups. The MVR (n = 2), and PVR + TVR (n = 2). The valves valve types included were mainly SJM and Medtronic implanted were St. Jude Medical (SJM; St. Paul, MN, Advantage; only two patients had received monocusp USA) or Medtronic Advantage (Minneapolis, MN, valves (Medtronic Hall, Minneapolis, MN, USA).
USA), both of which were bileaflet. All patients were Apart from their prosthetic heart valves, these patients provided with the Thrombocheck during their hospital were healthy. Ninety-four (92%) of the patients were stay for valve replacement surgery. Patients were INR self-managers. As part of a voluntary interview trained to handle the device in addition to conducting with regard to their quality of life, the patients were International Normalized Ratio (INR) self-manage- prepared to provide information about the handling ment using the ‘CoaguCheck’ device. The mean follow and user-friendliness of the device.
up of these patients was 3.2 ± 2.1 months. The resultswere double-blinded and decoded by an independent supervisor. Registration of the digital frequency analy-sis of the sound phenomena was encoded; this was Among the present patient cohort, 14 patients were transmitted every other day and decoded and con- found to have suspicious sound signals (3.8%), and trolled by a person unknown to either the patient or to these were confirmed in 13 cases by using alternative the study center. If the decoded signal demonstrated diagnostic methods. In 13 patients a pathological valve any suspicious sound phenomena, the study center function was confirmed by fluoroscopy, while in four cases neither transthoracic echocardiography (TTE) A detailed clinical follow up focusing especially on nor transesophageal echocardiography (TEE) revealed potential microemboli was performed every three a suspicious signal. A normal valve sound returned in months by means of patient interviews.
four patients after thrombolysis had been carried out.
Subgroup 2
Subgroup 1: (double-blind, prospective)
A multicenter study is currently investigating The patients managed the handling and transfer of whether anticoagulation therapy without coumarin the encoded data without any problems. Prosthetic derivatives after implantation of a mechanical valve valve endocarditis, peripheral or central embolizations prosthesis is reasonable. All 30 patients currently were not diagnosed during the entire observation peri- involved in the study were provided with a od. Among all of the INR measurements, 87% were ThromboCheck device and trained in its use while still in-patients after their valve surgery. In contrast to sub- One female patient transmitted a suspicious fre- group 1, the results of the digital frequency analysis quency spectrum for three consecutive days at approx- Home monitoring of prosthetic valves 659 Figure 1: A) Frequency spectrum of a prosthetic valve with Figure 2: A) Focal point of frequency before and after regular function. Note the characteristic opening and surgical removal of thrombus. B) Relationship of acoustic closure sounds of a bileaflet valve. B) Frequency spectrum energy of the two leaflets before and after surgical removal of a valve with impaired leaflet motion. Only the opening and closure sound of one leaflet is recorded. imately six weeks postoperatively. At the same time, two suspicious measurements, leading to a request to the INR values shifted between 1.7 and 2.1 (INR self- management). At one day after the INR had reachedthe target therapeutic range (≥2.8), the frequency spec- Subgroup 3
trum was normalized. As the patient remained com- Ten of 71 patients (11%) demonstrated a suspicious pletely asymptomatic, no further diagnostics were frequency analysis during routine check-up. In nine patients an impaired valve function could be demon- The vast majority of patients felt that the device was strated, and one suspicious frequency analysis was easy to use, reliable, and improved their well-being.
interpreted as a false negative. Four patients with one Following a special adaptation of the software, the leaflet stuck at the closed position were admitted for frequency analysis of patients with two prosthetic thrombolytic therapy with recombinant tissue plas- valves was found to be reliable, even if they were not minogen activator. At 48 h after thrombolysis, three patients again showed an inconspicuous frequencysignal (8).
Subgroup 2
In this subgroup, three patients (10%) presented with Subgroup 4
a pathological frequency analysis. The device recorded In this subgroup, neither clinical symptoms nor fre- 660 Home monitoring of prosthetic valves Table I: Reported patient experiences with the quency spectra indicative of prosthetic malfunctionwere detected during a mean observation period of 4.2 ± 1.3 months. The results of the questionnaire regard-ing user-friendliness, handling, technical problems Figure 3: A) Acoustic signal in the patient with and quality of life are listed in Table I.
undisturbed valve function shows two amplitude maxima representing the closure clicks of both leaflets. B) Acoustic Exemplary cases
signal with impaired leaflet motion; only one maximum is Case no. 1
In this 64-year-old male (who underwent AVR with a bileaflet, On-X, valve), anticoagulation therapy wasswitched from warfarin plus aspirin to aspirin alone.
After three months, the ThromboCheck recorded ic valve revealed massive thrombus formation at the pathological frequency spectra indicative of impaired valve prosthesis. Before and after surgery the valve leaflet motion for three consecutive days (Fig. 1A).
sounds were recorded with a ThromboCheck. Figure Neither TTE nor TEE identified any pathology.
2A and B illustrate the focal point of frequency and the Fluoroscopy confirmed valve dysfunction with incom- relationship of the acoustic energy of the two leaflets.
plete occluder motion of one leaflet. At four days after Both parameters were changed significantly.
the first pathological frequency signal had been docu- Subsequently, because of another operation, antico- mented, medication with heparin plus warfarin was agulation was switched to unfractionated heparin. At started. After five days, the ThromboCheck signals five days postoperatively, the ThromboCheck meas- ured a pathological frequency spectrum typical forrestricted leaflet motion (Fig. 3A and B). Neither TTE Case no. 2
nor TEE revealed any irregularity. X-ray fluoroscopy This 72-year-old female (AVR bileaflet valve, SJM), demonstrated impaired leaflet motion.
stopped oral anticoagulation prior to an orthopedicoperation. Several weeks later, the patient consulted Case no. 3
her physician because of dyspnea. TEE revealed an This 74-year-old male underwent AVR with a SJM impaired motion of both leaflets, without any demon- 31-mm composite valve graft, but was admitted to stration of thrombus. Surgical revision of the prosthet- hospital at 38 days postoperatively for bleeding from Home monitoring of prosthetic valves 661 esophageal varices due to liver cirrhosis. This patientwas monitored with the ThromboCheck once daily.
The first signal had already documented changes with-in the primary frequency spectrum. Invasive hemody-namic monitoring with a Swan-Ganz catheter wasinconspicuous. Consecutively performed TEE andangiography revealed complete immobilization of oneleaflet, while motion of the other leaflet was regardedas ‘normal’. However, the frequency spectrum showeda significant reduction in the amplitude of both leaflets(Fig. 4A-C).
Case no. 4
This 64-year-old female underwent AVR with an SJM valve. She was admitted to hospital with suspectedprosthetic valve dysfunction; this was diagnosed byTEE, and confirmed by investigation with theThromboCheck. On reoperation, thrombotic materialwas found predominantly at the sewing ring (Fig. 5).
Discussion
Previous investigations in animals and patients have demonstrated that the acoustic sound phenomena ofprosthetic heart valves remain constant over time.
However, the same prosthetic device, when implantedin various animals and patients, may produce differentsound pressures. These differences may be due to indi- vidual flow and resonance conditions. Furthermore, itcould be demonstrated that each valve replacementdevice produces characteristic frequency spectra.
Pathological processes involving heart valve pros- theses result in distinct changes to the acoustic signalsthat are characteristic of the respective device.
Previous experimental studies have shown that thespectral analysis of click sounds indicated prostheticvalve dysfunction long before echocardiography, fluo-roscopy or hemodynamic measurements were sugges-tive of any disturbed valve function. Thrombolysisdoes not seem justified if only the frequency analysis shows a pathological result, though the patient shouldbe intensely screened. Thus, a suitable hand-helddevice for signal recording represents an ideal option Figure 4: A) The course of ThromboCheck measurements for the home monitoring of prosthetic valve integrity.
shows a reduction in fingerprint similarity to less than It should be conceded that the frequency spectrum of 10%. B) The intact prosthetic valve’s frequency spectrum shows a maximum for each leaflet. The thrombosed prosthetic clicks must be influenced by changes in prosthetic valve’s frequency spectrum only shows a ‘single- hemodynamics or atrial fibrillation. A screening click’ because of complete blockage of one leaflet. The method requires high sensitivity but not necessarily amplitude was reduced by factor 30, representing the high specificity, as differential diagnosis will be made beginning of thrombus formation at the second leaflet. C) by the cardiologist once the control device has indicat- Fast-Fourier transformation (FFT) of the signal shows that certain frequencies are significantly reduced, or could no Fourteen patients of subgroups 1, 2 and 3 were iden- tified by pathological frequency analysis (1.1% ingroup 1; 10% in group 2; 11% in group 3). Only in onepatient in subgroup 3 was the frequency analysis inter- 662 Home monitoring of prosthetic valves application of innovative technology in their privateliving space and healthcare. They were also askedwhether they were prepared to use telemonitoring andsmart home care in their personal surroundings. Infact, 96% of the participants supported the idea ofusing such features and systems, especially for the careand monitoring of chronically ill people. In addition,83% expected that their desire for more safety wouldbe met, and 73.3% felt that this technology wouldprobably save time and increase the mobility and inde-pendence of older patients (70.8%). Only 11.7% sug-gested that telemonitoring and smart hometechnologies were unnecessary and would only causefears (18).
Faced with this background - and considering the fact that whilst home monitoring cannot substitute for Figure 5: A thrombosed prosthetic valve (St. Jude Medical); regular physician consultation, it might minimize time thrombotic material was found predominantly at the intervals between investigations and allow early diag- nosis of prosthetic valve-associated complications -patients may benefit significantly from devices which preted as a false negative. However, the subgroups offer them greater safety and an improved quality of included different patient populations (positive selec- tion regarding age, compliance, INR self-management, The efforts aimed at a maximum reduction in INR etc.). The present results suggest that prosthetic valvu- levels - and therefore at a further reduction in hemor- lar dysfunction without clinically apparent symptoms rhage-related complications - seem reasonable. It even seems to occur more frequently than is generally pre- seems realistic that in the near future, and under cer- sumed. This suggestion is supported by the results of tain conditions, a patient who has undergone heart studies dealing with the occurrence of high-intensity valve replacement with a mechanical prosthesis can be treated without the use of coumarin derivatives. One This observation could be of imminent importance, of these conditions would be the availability of a tool especially in developing countries where a high per- such as the ThromboCheck, which is able to analyze centage of the inhabitants live in rural areas and there the frequency spectra of valve sounds. Whether in the are huge distances between medical centers, with con- future patients bearing mechanical prostheses may be sequent long time intervals between INR measure- treated solely with aspirin, for example, remains to be ments. The ThromboCheck provides reproducible seen. However, the pursuit of this path - which would results of frequency analyses, and is also easy to use.
certainly lead to an elimination of hemorrhage-related Moreover, prosthetic valve dysfunction is detected complications - requires that clotting-related complica- early enough to prevent severe, or even fatal, conse- tions are detected immediately and treated when nec- Several groups have studied the acoustic signals of Björk-Shiley mechanical valves following strut fracture Therapeutic considerations
(10-17), but none of the reports of in-vivo testings has The incidence of prosthetic heart valve obstruction presented individual baseline measurements revealing through non-infectious mass is considered to be 0.5% an individual, fingerprint-like frequency spectrum.
per year after AVR, and up to 2.5% per year after MVR The present authors’ studies have contributed to the (19). These values are considerably higher in poorly development of the first hand-held device for home controlled populations, for example in developing monitoring of heart valve function, and this is current- countries. Several studies have indicated the impact of ly being tested under clinical conditions.
thrombolysis or valve re-replacement on prognosis One potential disadvantage of such a device is an (20-23). Recently, the present authors presented pre- increase in patient concern. However, the results of the liminary results of simulation and animal studies relat- questionnaire indicated that patients may feel safer, ed to the recording and transmission of acoustic sound and that the device may help them to cope with the phenomena (5). Further investigations have concen- inevitable disadvantages of a mechanical heart valve.
trated on a reliable implementation of the Doppler In a recent study, 307 persons aged 50 years or more probe. If these efforts progress successfully, it could be were interviewed with regard to their attitude to the envisaged that this type of home monitoring will be Home monitoring of prosthetic valves 663 made available to a large patient clientele. The present DW, Stein PD. Low frequency analysis of opening type of device will certainly be important in future for sound for detection of single leg separation of the early detection of relevant dysfunctions in heart Björk-Shiley convexo-concave heart valves. J Heart valve prostheses, as well as in the observation of med- ically treated cardiac patients, a reduction in their anti- 11. Candy N, Jones HE. Classification of prosthetic heart valve sounds: A parametric approach. J thromboembolic and bleeding complications.
12. Walker DK, Scotten LN. Discrimination in vitro References
between the acoustic emissions from Björk-Shiley 1. Horstkotte D, Piper C, Schulte HD.
convexo-concave valves with and without a broken Thrombosierung prosthetischer Herzklappen: minor strut. Med Biol Eng Comput 1991;29:457-464 13. Candy N, Jones HE. Processing of prosthetic heart valve sounds for single leg separation classification.
2. Jamieson WR, Munro AI, Miyagishima RT, et al.
Multiple mechanical valve replacement surgery 14. Dow JJ, Plemons TD, Scarbrough K, et al. Acoustic comparison of St. Jude Medical and CarboMedics assessment of the physical integrity of Björk-Shiley prostheses. Eur J Cardiothorac Surg 1998;13:151-159 convexo-concave heart valves. Circulation 3. Munclinger MJ, Patel JJ, Mitha AS. Thrombolysis of thrombosed St. Jude Medical prosthetic valves: 15. Kien GA, Jeffries B, Katz HV, et al. Digital acousti- Rethrombosis - a sign of tissue ingrowth. J Thorac cal analysis of normal and bimodal Björk-Shiley 60° convexo-concave heart valves. Am J Cardiol 4. Alexiou C, McDonald A, Langley SM, Dalrymple- Hay MJ, Haw MP, Monro JL. Aortic valve replace- 16. Eberhardt A, Chassaing C, Ward M, Lewandowski ment in children: Are mechanical prostheses a good S. Acoustical characterization of mechanical valve option? Eur J Cardiothorac Surg 2000;17:125-133 condition and loading. J Heart Valve Dis 1995;4:649- 5. Eitz T, Fritzsche D, Grimmig O, et al. Acoustic phe- nomena and valve dysfunction in cardiac prosthe- 17. Plemons TD, Hovenga M. Acoustic classification of ses: Data acquisition and collection via the internet.
the state of artificial heart valves. J Acoust Soc Am 6. Fritzsche D, Grimmig O, Eitz T, Brensing A, 18. Böhm U, Röhrig A, Schadow B. Telemonitoring and Minami K, Koerfer R. Home monitoring of patients smart home care. Hohe Akzeptanz bei den über 50- after prosthetic valve replacement. Experimental Jährigen. Deutsches Ärzteblatt 2003;50:2743-2744 background and first clinical attempts. Thorac 19. Horstkotte D, Pieper C, Schulte HD.
Thrombosierung prosthetischer Herzklappen: 7. Klein RM, Niehues R, Horstkotte D, Schulte H-D, Diagnostik und Management. Z Kardiol 1998;87:20- dysfunktion nach Bügelbruch von konvex-konkav 20. Shapira Y, Herz I, Birnbaum Y, Snir E, Vidne B, Björk-Shiley Monoklappenscheibenprothesen; Sagie A. Repeated thrombolysis in prosthetic heart Häufigkeit, Diagnostik und Management.
valves: A report of three cases and review of the lit- Intensivmedizin und Notfallmedizin 1995;32:462- erature. J Heart Valve Dis 2000;9:146-149 21. Birdi I, Angelini G, Bryan A. Thrombolytic therapy 8. Ben Zekry S, Sagie A, Ben-Dor I, Nukrian H, Battler for left sided prosthetic heart valve thrombosis. J A, Shapira Y. Initial clinical experience with hand- held device (Thrombocheck®) for the detection of 22. Lengyel M, Vegh G, Vandor L. Thrombolysis is bileaflet prosthetic valve malfunction. J Heart Valve superior to heparin for non-obstructive mitral mechanical valve thrombosis. J Heart Valve Dis 9. Hiratsuka R, Fukunaga S, Tayama E, et al. High- intensity transient signals due to prosthetic valve 23. Renzulli A, Vitale N, Caruso A, Dialetto G, Schinosa obstruction: Diagnostic and therapeutic implica- L, Cotrufo M. Thrombolysis for prosthetic valve thrombosis: Indications and results. J Heart Valve 10. Durand LG, Grenier MC, Inderbitzen R, Wieting

Source: https://mijn.denationaletrombosedienst.nl/downloads/14415_Fritzsche_2874.pdf