Medical system for detecting atrial tachyarrhythmias by classification based on heart rate variability

Abstract

A medical device (60) for detecting atrial tachyarrhythmia comprises: a processor configured to determine RR intervals between consecutive R waves of a cardiac electrical signal, and to determine a classification factor from R waves identified within a predetermined time period by determining at least a first classification factor associated with variability of the RR intervals and a second classification factor indicative of the presence of a ventricular tachyarrhythmia. The processor is configured to classify the cardiac electrical signal for the predetermined time period as unclassified, atrial tachyarrhythmia, or non-atrial tachyarrhythmia based on a comparison of the determined classification factor with a classification criterion.

Description

[0001] This application is a divisional application of an invention patent application with an international application date of 2017-03-06, international application number of PCT / US2017 / 020877, application number of 201780022408.0 entering the Chinese national phase, and titled "Medical system for detecting atrial tachyarrhythmia through classification based on heart rate variability". Technical Field

[0002] The present disclosure relates generally to cardiac medical devices, and in particular to cardiac medical devices and methods for detecting atrial arrhythmia episodes from sensed cardiac electrical signals. Background Art

[0003] During normal sinus rhythm (NSR), the heartbeat is regulated by electrical signals generated by the sinoatrial (SA) node located in the wall of the right atrium. Each atrial depolarization signal generated by the SA node propagates across the atria (causing depolarization and contraction of the atria) and arrives at the atrioventricular (AV) node. The AV node responds by propagating the ventricular depolarization signal through the bundle of His of the ventricular septum and then to the bundle branches and Purkinje muscle fibers of the right and left ventricles.

[0004] Atrial tachyarrhythmias include unorganized forms of atrial fibrillation and various degrees of organized atrial tachycardia, including atrial flutter. Atrial fibrillation (AF) occurs due to multiple focal triggers in the atria or due to changes in the atrial matrix causing non-uniformity of conduction through different atrial regions. Ectopic triggering may originate anywhere in the left atrium or right atrium or pulmonary veins. The AV node will be bombarded by frequent and irregular atrial activation, but will only conduct depolarization signals when the AV node is not in the refractory period. The ventricular cycle length will be irregular and will depend on different states of the refractory of the AV node.

[0005] As the more serious consequences of sustained atrial arrhythmias have become known, such as relatively more severe ventricular arrhythmias and the associated risk of stroke, there has been increasing interest in monitoring and treating atrial arrhythmias. Implantable cardiac monitors and implantable cardioverter-defibrillators (ICDs) can be configured to acquire cardiac electrical signals that can be analyzed to detect atrial arrhythmias. Summary of the invention

[0006] In general, the present disclosure relates to techniques for detecting cardiac events, and more specifically, atrial tachyarrhythmia episodes, by a medical device. A medical device operating according to the techniques disclosed herein analyzes cardiac electrical signals over a plurality of time periods and classifies each of the time periods based on characteristics of the cardiac electrical signals, such as characteristics of the RR intervals occurring during each of the plurality of time periods, and whether there are other factors that cause the time period to be unclassified, such as whether other tachyarrhythmia episodes are being detected, among others. When a predetermined number of time periods are classified as atrial tachyarrhythmias, an atrial tachyarrhythmia may be detected, and the device provides a response to the detection by storing the episodes of the cardiac electrical signals.

[0007] In one example, the present disclosure provides a method for detecting atrial tachyarrhythmia by sensing a cardiac electrical signal from a first sensing electrode vector, identifying an R wave in the cardiac electrical signal that accompanies ventricular depolarization, determining an RR interval between consecutive R waves in the first cardiac electrical signal, and determining a classification factor from the R waves identified within a predetermined time period by at least determining a first classification factor associated with the variability of the RR interval and a second classification factor indicating the presence of a ventricular tachyarrhythmia. The method further includes classifying the predetermined time period as one of unclassified, atrial tachyarrhythmia, and non-atrial tachyarrhythmia by comparing the determined classification factor with a classification criterion, wherein the predetermined time period is classified as unclassified in response to the second classification factor indicating the presence of a ventricular tachyarrhythmia. The method detects an atrial tachyarrhythmia episode in response to the first time period being classified as unclassified and at least one previous time period and at least one subsequent time period being classified as an atrial tachyarrhythmia. In response to detecting an atrial tachyarrhythmia episode, episode data determined from at least the first cardiac electrical signal is stored in a memory of the medical device for transmission to an external device.

[0008] In another example, the present disclosure provides a medical device for detecting atrial tachyarrhythmia. The medical device includes a sensing circuit configured to sense a cardiac electrical signal from a sensing electrode vector, a processor, and a memory accessible to the processor. The processor is configured to: identify an R wave associated with ventricular depolarization in a first cardiac electrical signal; determine an RR interval between consecutive R waves in the first cardiac electrical signal; determine a classification factor from R waves identified within a first predetermined time period by at least determining a first classification factor associated with the variability of the RR interval and a second classification factor indicating the presence of a ventricular tachyarrhythmia; classify the first predetermined time period into one of unclassified, atrial tachyarrhythmia, and non-atrial tachyarrhythmia by comparing the determined classification factor with a classification criterion, wherein the first predetermined time period is classified as unclassified in response to the second classification factor indicating the presence of a ventricular tachyarrhythmia; detect an atrial tachyarrhythmia episode in response to the first time period being classified as unclassified and at least one previous time period and at least one subsequent time period being classified as an atrial tachyarrhythmia; and in response to detecting an atrial tachyarrhythmia episode, store episode data determined from at least the first cardiac electrical signal in a memory for transmission to an external device.

[0009] In another example, the present disclosure provides a non-transitory computer-readable storage medium storing instructions for causing a processor included in a medical device to perform a method for detecting an atrial tachyarrhythmia. The method includes: sensing a first cardiac electrical signal from a first sensing electrode vector; identifying an R wave in the first cardiac electrical signal that is accompanied by ventricular depolarization; determining the RR interval between consecutive R waves in the first cardiac electrical signal; determining a classification factor from R waves identified within a first predetermined time period by at least determining a first classification factor associated with the variability of the RR interval and a second classification factor indicating the presence of ventricular tachyarrhythmia; classifying the first predetermined time period into one of unclassified, atrial tachyarrhythmia and non-atrial tachyarrhythmia by comparing the determined classification factor with a classification criterion, wherein the first predetermined time period is classified as unclassified in response to the second classification factor indicating the presence of ventricular tachyarrhythmia; detecting an atrial tachyarrhythmia attack in response to the first time period being classified as unclassified and at least one previous time period and at least one subsequent time period being classified as atrial tachyarrhythmia; and storing the attack data determined from at least the first cardiac electrical signal in a memory of the medical device for transmission to an external device in response to detecting the atrial tachyarrhythmia attack.

[0010] This summary is intended to provide an overview of the subject matter described in this disclosure. It is not intended to provide an exclusive or exhaustive explanation of the devices and methods described in detail in the following figures and description. Further details of one or more examples are set forth in the following figures and description. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1A is a conceptual diagram of an implantable medical device (IMD) system for detecting atrial arrhythmias according to one example.

[0012] Figure 1B is a conceptual diagram of an IMD system for detecting atrial tachyarrhythmias according to another example.

[0013] Figure 1C is a conceptual diagram of yet another IMD system in which the techniques disclosed herein for detecting atrial tachyarrhythmias may be implemented.

[0014] Figure 2 is a functional schematic diagram of the implantable cardioverter defibrillator (ICD) of FIG. 1 .

[0015] Figure 3A and Figure 3B is a conceptual diagram of an alternative ICD system that may be configured to detect atrial tachyarrhythmias according to the techniques disclosed herein.

[0016] Figure 4 is based on an example used by Figure 1A , Figure 1B and Figure 2 Any of the ICDs or Figure 1C Schematic diagram of a method of using a monitoring device to detect cardiac events.

[0017] Figure 5 is a graphical representation of a two-dimensional histogram representing the region of a Lorenz plot for detecting atrial tachyarrhythmias used in the techniques disclosed herein.

[0018] Figure 6 is a flow chart of a method for determining a factor for classifying a time period for detecting an atrial tachyarrhythmia according to one example.

[0019] Figure 7 is a flow chart of a method for classifying a predetermined time period for detecting an atrial tachyarrhythmia according to one example.

[0020] Figure 8 It can be done by Figure 1A , Figure 1B and Figure 3A The ICD shown in Figure 1CSchematic diagram of atrial fibrillation detection performed by a monitor.

[0021] Fig. 9 is a schematic diagram of a method for detecting atrial fibrillation by an ICD or implantable monitoring device according to another example.

[0022] Fig.10 is a flow chart of a method for detecting atrial fibrillation according to one example.

[0023] Fig.11 is a flow chart of a method performed by an ICD or implantable monitor for providing a response to detection of an atrial tachyarrhythmia, according to one example. DETAILED DESCRIPTION

[0024] In the following description, reference is made to illustrative embodiments for implementing the methods described herein. It should be understood that other embodiments may be used without departing from the scope of the present disclosure.

[0025] In various examples, a cardiac electrical signal is used to determine consecutive ventricular cycle lengths for detecting atrial arrhythmias. The ventricular cycle length can be determined as the interval between consecutive R waves sensed from the cardiac electrical signal and accompanied by ventricular depolarization. The difference between consecutive RR intervals (RRI) is analyzed for evidence of atrial tachyarrhythmias (e.g., atrial fibrillation). As described herein, time periods of a cardiac signal can be classified as AF, non-AF, or unclassified based on analysis of the RRI and other factors. When a predetermined number of time periods of a cardiac signal are classified as AF, a medical device operating according to the techniques disclosed herein can detect AF. However, the device can adjust the classification criteria applied to classify time periods of the cardiac signal before detecting AF, and detect AF based on the adjusted classification criteria applied to classify subsequent time periods.

[0026] Aspects of the methods described herein can be incorporated in various implantable medical devices or external medical devices with cardiac signal monitoring capabilities, which may or may not include treatment delivery capabilities. Such devices include single-chamber, dual-chamber or biventricular pacing systems or ICDs that sense R waves and deliver electrical stimulation therapy to the ventricles. The atrial arrhythmia detection method disclosed herein can also be incorporated into an implantable cardiac monitor with implantable electrodes or an external cardiac monitor (e.g., Holter monitor) with electrocardiogram (ECG) electrodes coupled to the patient's skin to detect R waves or incorporated in a computerized system that analyzes pre-recorded ECG and electrocardiogram (EGM) data. Each embodiment can be further implemented in a patient monitoring system, such as a centralized computer system that processes cardiac electrical signals and other data sent to it by an implantable or wearable monitoring device.

[0027] FIG1 is a conceptual diagram of an implantable medical device (IMD) system 1 for detecting atrial arrhythmias according to one example. The IMD system 1 of FIG1 includes an implantable cardioverter-defibrillator (ICD) 10 coupled to a patient's heart 2 via transvenous electrical leads 6, 11, and 16. The ICD 10 includes a connector block 12 that can be configured to receive the proximal ends of a right ventricle (RV) lead 16, a right atrium (RA) lead 11, and a coronary sinus (CS) lead 6 that are advanced transvenously for positioning electrodes for sensing and stimulation in three or all four heart chambers.

[0028] RV lead 16 is positioned so that its distal end is in the right ventricle for sensing RV heart signals and delivering pacing or shock pulses in the right ventricle. For these purposes, RV lead 16 is equipped with pacing and sensing electrodes, shown as ring electrode 30 and tip electrode 28. In some examples, tip electrode 28 is an extendable spiral electrode retractably mounted within electrode head 29. RV lead 16 is further shown as carrying defibrillation electrodes 24 and 26, which can be elongated coil electrodes for delivering high voltage cardioversion / defibrillation (CV / DF) electrodes. Because defibrillation electrode 24 can be carried along RV lead 16, so that when distal pacing and sensing electrodes 28 and 30 are positioned for pacing and sensing in the right ventricle, they are substantially positioned in the right ventricle, so defibrillation electrode 24 is referred to herein as "RV defibrillation electrode" or "RV coil electrode". Because the defibrillation electrode 26 can be carried along the RV lead 16 so that when the distal end of the RV lead 16 is advanced into the right ventricle, it is at least partially positioned along the superior vena cava (SVC), the defibrillation electrode 26 is referred to herein as a "SVC defibrillation electrode" or "SVC coil electrode."

[0029] Each of electrodes 24, 26, 28, and 30 is connected to a respective insulated conductor extending within the body of lead 16. The proximal ends of the insulated conductors are coupled to corresponding connectors carried by proximal lead connector 14 (e.g., a DF-4 connector) at the proximal end of lead 16 for providing an electrical connection to ICD 10. It is to be understood that although ICD 10 is shown in FIG. 1 as a multi-chamber device coupled to RA lead 11 and CS lead 6, ICD 10 may be configured as a single-chamber device coupled only to RV lead 16, as described below in conjunction with Figure 1B As shown and described. The techniques disclosed herein for detecting atrial tachyarrhythmias can be successfully performed without the need for atrial signal sensing. Therefore, in some examples, RA lead 11 is optional.

[0030] If RA lead 11 is included, RA lead 11 can be positioned so that its distal end is in the vicinity of the right atrium and superior vena cava. Lead 11 is equipped with pacing and sensing electrodes 17 and 21, shown as tip electrode 17 (which can be an extendable helical electrode retractably mounted within electrode head 19) and ring electrode 21 spaced proximally from tip electrode 17. Electrodes 17 and 21 provide sensing and pacing in the right atrium and are each connected to a corresponding insulated conductor within the body of RA lead 11. Each insulated conductor is coupled at its proximal end to a connector carried by proximal lead connector 13.

[0031] CS lead 6 is also optional and is not required for successful performance of the atrial tachyarrhythmia detection methods disclosed herein. When present, CS lead 6 can be advanced within the vasculature of the left side of the heart via the coronary sinus and cardiac vein 18. Figure 1A In the embodiment of the present invention, the CS lead 6 is shown as having one or more electrodes 8, which can be used in conjunction with the RV coil electrode 24 or the SVC coil electrode 26 to deliver an electric shock for cardioversion and defibrillation therapy. In other examples, the coronary sinus lead 6 can also be equipped with one or more electrodes 8 for delivering pacing and or sensing cardiac electrical signals in the left chamber of the heart (i.e., the left ventricle and / or the left atrium). The one or more electrodes 8 are coupled to corresponding insulated conductors within the body of the CS lead 6, which provides a connection to the proximal connector 4.

[0032] RV pacing and sensing electrodes 28 and 30 may be used as a bipolar pair, commonly referred to as a "tip-to-ring" configuration for sensing cardiac electrical signals. Further, RV tip electrode 28 may be selected with coil electrode 8, 24, or 26 for use as an integrated bipolar pair, commonly referred to as a "tip-to-coil" configuration for sensing cardiac electrical signals. ICD 10 may, for example, select one or more sensing electrode vectors, including a tip-to-ring sensing vector between electrodes 28 and 30 and a tip-to-coil sensing vector, for example, between RV tip electrode 28 and SVC coil electrode 26, between RV tip electrode 28 and RV coil electrode 24, between RV ring electrode 30 and SVC coil electrode 26, or between RV ring electrode 30 and RV coil electrode 24. In some cases, any of electrodes 24, 26, 28, or 30 carried by RV lead 16 may be selected by ICD 10 to be in a unipolar sensing configuration, with ICD housing 15 acting as a neutral electrode, commonly referred to as a "can" or "shell" electrode. It should be appreciated that numerous sensing and electrical stimulation electrode vectors using individual electrodes carried by one or more of leads 6, 11, and 16 coupled to the ICD 10 may be available, and that the ICD 10 may be configured to selectively couple one or more sensing electrode vectors to a sensing circuit system enclosed by the housing 15, such as a sensing circuit system including one or more amplifiers, filters, rectifiers, comparators, sensing amplifiers, analog-to-digital converters, and / or other circuit systems configured to acquire cardiac electrical signals for detecting arrhythmias.

[0033] In other examples, ICD housing 15 may function as a subcutaneous defibrillation electrode in conjunction with one or more of coil electrodes 8, 24, or 26 for delivering CV / DF shocks to the atria or ventricles. It should be appreciated that alternative lead systems may be substituted for Figure 1A The three-lead system shown in Figure 1A A specific multi-chamber ICD and lead system are shown in the present invention, but the methods included in the present invention can be adapted for use with any single-chamber ICD, dual-chamber ICD or multi-chamber ICD or pacemaker system, subcutaneously implantable device, or other internal or external cardiac monitoring device.

[0034] External device 40 is shown in telemetric communication with ICD 10 via RF communication link 42. External device 40 is often referred to as a "programmer" because it is typically used by a physician, technician, nurse, clinician, or other qualified user to program operating parameters in ICD 10. External device 40 may be located in a clinic, hospital, or other medical facility. External device 40 may alternatively be embodied as a home monitor or handheld device that may be used in a medical facility, in a patient's home, or in another location. Operating parameters such as sensing and therapy delivery control parameters may be programmed into ICD 10 using external device 40.

[0035] External device 40 includes processor 52, memory 53, user display 54, user interface 56, and telemetry circuit system 58. Processor 52 controls the operation of the external device and processes data and signals received from ICD 10. According to the technology disclosed herein, processor 52 receives sensing vector data obtained by ICD 10 and transmitted from ICD 10 to telemetry circuit system 58. As shown below in conjunction with Fig.11 As described, the ICD 10 may be configured to store cardiac signal data associated with a detected atrial tachyarrhythmia episode and transmit the cardiac signal data to an external device 40. The processor 52 provides at least a portion of the cardiac electrical signal data to a user display 54 for use in generating a display of the cardiac electrical signal detected as an atrial tachyarrhythmia for observation and review by a clinician.

[0036] The user display 54 provides a display of cardiac signal data and may include a graphical user interface that facilitates programming of one or more sensing parameters and / or atrial arrhythmia detection parameters by user interaction with the external device 40. The external device 40 may display other data and information related to ICD functionality to the user for viewing ICD operation and programmed parameters as well as cardiac electrical signals or other physiological data retrieved from the ICD 10 during an interrogation session. The user interface 56 may include a mouse, touch screen, or other pointing device, keyboard, and / or keypad to enable the user to interact with the external device 40 to initiate a telemetry session with the ICD 10 for retrieving data from and / or transmitting data to the ICD 10, and for selecting and programming desired sensing and therapy delivery control parameters into the ICD 10.

[0037] Telemetry circuitry 58 includes a transceiver and antenna configured for bidirectional communication with an implanted transceiver and antenna included in ICD 10. Telemetry circuitry 58 is configured to operate in conjunction with processor 52 for encoding and decoding transmitted and received data related to ICD functions via communication link 42. A radio frequency (RF) link (such as Bluetooth) may be used. The external device 40 may be configured to establish a communication link 42 between the ICD 10 and the external device 40 using a Wi-Fi, Medical Implant Communications Service (MICS), or other RF bandwidth. In some examples, the external device 40 may include a programming header that is placed adjacent to the ICD 10 to establish and maintain the communication link, and in other examples, the external device 40 and ICD 10 may be configured to communicate using a range telemetry algorithm and circuitry that does not require the use of a programming header and does not require user intervention to maintain the communication link.

[0038] It is contemplated that the external device 40 may be connected to a communication network, either wired or wirelessly, via the telemetry circuitry 58 for transmitting data to a remote database or computer to allow remote management of the patient. A remote patient management system may be configured to utilize the techniques of the present disclosure to enable a clinician to review cardiac electrical signal data and atrial tachyarrhythmia episode data received from the ICD 10, and to select and program control parameters that are transmitted to the ICD 10. Reference is made to commonly assigned U.S. Patent Nos. 6,599,250 (Webb et al.), 6,442,433 (Linberg et al.), 6,418,346 (Nelson et al.), and 6,480,745 (Nelson et al.) for a general description and examples of remote patient management systems that enable remote patient monitoring and device programming.

[0039] Figure 1Bis a conceptual diagram of a single-chamber ICD 10' coupled to an RV lead 16. The techniques disclosed herein may be implemented in a single-chamber ICD that is coupled only to a ventricular lead, such as the RV lead 16, for receiving cardiac electrical signals including at least an R-wave accompanying ventricular depolarization of the heart 2. Electrodes 28 and 30 (and / or coil electrodes 24 and 26) may be used to acquire cardiac electrical signals required to perform atrial tachyarrhythmia detection as described herein without the need for an atrial sensing and pacing lead 11 as shown in FIG. 1 . The R-waves sensed from the cardiac electrical signals obtained by the ICD 10' are used to determine the RR interval (RRI) between consecutive sensed R-waves for detection of atrial tachyarrhythmias by a processor of the ICD 10' based at least in part on analysis of the RRI. The single chamber ICD 10′ may be configured to sense cardiac electrical signals from electrodes 24, 26, 28, and / or 30, detect atrial tachyarrhythmias, and provide an atrial tachyarrhythmia detection response, such as storing atrial tachyarrhythmia episode data for transmission to an external device 40 ( Figure 1A ). Single chamber ICD 10' may additionally be configured to deliver ventricular bradycardia pacing, detect ventricular tachyarrhythmias, and deliver anti-tachycardia pacing therapy and cardioversion / defibrillation shock therapy to the RV via electrodes 24, 26, 28, and / or 30 carried by leads 16.

[0040] Figure 1C is a conceptual diagram of a cardiac monitoring device 60 that can employ various aspects of the atrial tachyarrhythmia detection technology disclosed herein. The monitoring device 60 is shown as being implanted subcutaneously in the upper chest region of the patient's body 3 and offset from the patient's heart 2. The housing 62 of the cardiac monitor 60 (shown scaled up compared to the patient's body 3) includes a non-conductive head module 64 attached to the hermetically sealed housing 62. The housing 62 contains the circuitry of the cardiac monitor 60 and is typically conductive, but may be partially covered by an electrically insulating coating. A first subcutaneous sensing electrode A is formed on a surface of the head module 64, and a second subcutaneous sensing electrode B is formed by at least a portion of the housing 62. For example, when the housing 62 is coated with an electrically insulating coating, electrode B may be an exposed portion of the housing 62. The conductive housing electrode B may be directly connected to the sensing circuitry.

[0041] An electrical feedthrough extends through the mating surfaces of the head module 64 and the housing 62 to electrically connect the first sensing electrode A with the sensing circuitry enclosed within the housing 62. The electrical signals accompanying the depolarization and repolarization of the heart 2 are referred to as cardiac electrical signals and are sensed across the sensing electrodes A and B and include at least the R wave accompanying the ventricular depolarization of the heart 2. The cardiac monitoring device 60 can be sutured to the subcutaneous tissue to the axis of the heart 8 in the desired orientation of its electrodes A and B to detect and record the cardiac electrical signals in the sensing vector AB for subsequent processing and uplink telemetry transmission to the external device 40 ( Figure 1A ).

[0042] In one embodiment, the spacing between electrode A and electrode B can range from 60 mm to 25 mm. In other embodiments, the electrode spacing can range from 55 mm to 30 mm, or from 55 mm to 35 mm. The volume of the implantable cardiac monitoring device 60 can be three cubic centimeters or less, 1.5 cubic centimeters or less, or any volume between three cubic centimeters and 1.5 cubic centimeters. The length of the cardiac monitoring device 60 can range from 30 mm to 70 mm, 40 mm to 60 mm, or 45 mm to 60 mm, and can be any length between 30 mm and 70 mm. The width of the main surface of such a cardiac monitoring device 60 can range from 3 mm to 10 mm, and can be any thickness between 3 mm and 10 mm. The thickness of the cardiac monitoring device 60 can range from 2 mm to 9 mm, or from 2 mm to 5 mm, and can be any thickness between 2 mm and 9 mm.

[0043] The sensing circuit system included in the housing 62 is configured to detect R waves for monitoring atrial tachyarrhythmias according to the techniques disclosed herein. Such sensing circuit systems may include pre-filters and amplifiers, rectifiers, sensing amplifiers, analog-to-digital filters, comparators, and / or other components configured to receive cardiac electrical signals. Various aspects of cardiac monitoring devices of the type that may employ the atrial arrhythmia detection techniques disclosed herein are generally disclosed in U.S. Publication Nos. 2015 / 0088216 (Gordon et al.) and U.S. Pat. No. 7,027,858 (Cao et al.).

[0044] Typically, the hermetically sealed housing 62 includes a lithium battery or other power source, a processor and memory or other control circuitry that controls the operation of the device and records arrhythmia cardiac electrical signal episode data in memory registers, and a telemetry transceiver antenna and circuitry that receives downlink telemetry commands from the external device 40 and transmits the stored data to the external device 40 in a telemetry uplink. The circuitry and memory may be implemented in discrete logic or a microcomputer-based system with A / D conversion of sampled cardiac electrical signal amplitude values. An implantable cardiac monitor that may be modified in accordance with the techniques of the present disclosure is described in U.S. Patent No. 6,412,490 (Lee et al.).

[0045] Figure 2 is a functional schematic diagram of an ICD such as ICD 10 of Figure 1. This diagram should be considered illustrative of the types of devices that may be used to embody the techniques disclosed herein and is not intended to be limiting. Figure 2 The examples shown in are processor controlled devices, but the disclosed methods may also be practiced using other types of devices (such as devices using dedicated digital circuit systems). In other words, the processor 224 may include any combination of integrated circuit systems, discrete logic circuit systems, analog circuit systems, such as one or more microprocessors, digital signal processors (DSPs), application specific integrated circuits (ASICs), or field programmable gate arrays (FPGAs). In some examples, the processor 224 may include multiple components, such as any combination of the following: one or more microprocessors, one or more DSPs, one or more ASICs, or one or more FPGAs and other discrete or integrated logic circuit systems, and / or analog circuit systems.

[0046] for Figure 1A , the ICD 10 is provided with a plurality of connection terminals for achieving electrical connection to the leads 6, 11 and 16 and their respective electrodes. During monopolar stimulation or sensing, the housing 15 can be used as a neutral electrode. The electrodes 24, 26 and 8 can be selectively coupled to a high voltage output circuit 234 to facilitate the use of one or more of the coil electrodes 8, 24 and 26 and (optionally) the housing 15 to deliver high energy shock pulses to the heart.

[0047] The RA tip electrode 17 and the RA ring electrode 21 can be coupled to the atrial sensing amplifier 204 for sensing atrial signals such as P waves. The RV tip electrode 28 and the RV ring electrode 30 can be coupled to the ventricular sensing amplifier 200 for sensing ventricular signals. The atrial sensing amplifier 204 and the ventricular sensing amplifier 200 can take the form of automatic gain controlled amplifiers with adjustable sensitivity. The ICD 10, and more specifically, the processor 224 can automatically adjust the sensitivity of the atrial sensing amplifier 204, the ventricular sensing amplifier 200, or both in response to the detection of oversensing so as to reduce the possibility of oversensing of cardiac events and / or non-cardiac murmurs.

[0048] The atrial sense amplifier 204 and the ventricular sense amplifier 200 may receive timing information from the pacemaker timing and control circuitry 212. For example, the atrial sense amplifier 204 and the ventricular sense amplifier 200 may receive blanking period inputs (e.g., A_BLANK and V_BLANK, respectively) that indicate the amount of time to "turn off" the amplifier to prevent saturation due to an applied pacing pulse or defibrillation shock. The general operation of the ventricular sense amplifier 200 and the atrial sense amplifier 204 may correspond to the general operation disclosed in U.S. Pat. No. 5,117,824 (Keimel et al.). Whenever the signal received by the atrial sense amplifier 204 exceeds the atrial sensitivity, a signal is generated on the P-out signal line 206. Whenever the signal received by the ventricular sense amplifier 200 exceeds the ventricular sensitivity, a signal is generated on the R-out signal line 202. As described below, the signal on the R-out signal line 202, which may be referred to as a ventricular sensed event (Vs event) signal, may be received by the processor 224 and used to determine an RRI difference.

[0049] The switch matrix 208 is used to select which of the available electrodes 8, 17, 21, 24, 26, 28, and 30 are coupled to the broadband amplifier 210 for digital signal analysis. The selection of electrodes is controlled by the processor 224 via the data / address bus 218. The selected electrode configuration may be changed as desired for the various sensing, pacing, cardioversion, and defibrillation functions of the ICD 10. For example, while the RV electrodes 28 and 30 are shown as being coupled to the sense amplifier 200 and pacing output circuit 216 suggesting dedicated pacing / sense electrodes, and the coil electrodes 24 and 26 are shown as being coupled to the HV output circuit 234 suggesting dedicated CV / DV shock electrodes, it should be appreciated that the switch circuitry included in the switch matrix 208 may be used to select any of the available electrodes in the sensing electrode vector, the pacing electrode vector, or the CV / DF shock vector as previously indicated.

[0050] Signals from the electrodes selected for coupling to the bandpass amplifier 210 are provided to the multiplexer 220 and thereafter converted by the A / D converter 222 to a multi-bit digital signal for storage in the memory 226 via the data / address bus 218 under the control of the direct memory access circuit 228. The processor 224 may employ digital signal analysis techniques to characterize the digitized signals stored in the memory 226 to identify and classify the patient's cardiac rhythm using any of a number of signal processing methods for analyzing cardiac signals and cardiac event waveforms (e.g., P and R waves). A tachyarrhythmia detection system is described in U.S. Pat. No. 5,545,186 (Olson et al.).

[0051] It should be understood that the Figure 2 For example, Figure 1B The single-chamber ICD 10' may include a ventricular sensing amplifier 200 and a ventricular pacing output circuit 216 and terminals for electrically coupling to electrodes 24, 26, 28, and 30, while an atrial sensing amplifier 204, an atrial pacing output circuit 214 and terminals for electrically coupling to electrodes 8, 17, and 21 may be omitted and / or coupled to other electrodes. For example, the sensing amplifier 204 and / or the pacing output circuit 214 may be coupled to electrodes 24, 26, and / or the housing electrode 15.

[0052] When an arrhythmia is detected, the cardiac attack signal data can be stored in the memory 226 along with the sensed intervals and corresponding annotations of the sensed events. The cardiac electrical signals sensed from the programmed sensing electrode pairs can be stored as EGM signals. Typically, the near-field sensing electrode pairs include a tip electrode and a ring electrode located in the atrium or ventricle, such as RA electrodes 17 and 21 or RA electrodes 28 and 30. The far-field sensing electrode pairs include electrodes that are spaced farther apart, such as any of the following pairs: defibrillator coil electrode 8, 24 or 26 and housing 15; tip electrode 17 or 28 and housing 15; tip electrode 17 or 28 and defibrillator coil electrode 8, 24 or 26; or atrial tip electrode 17 and ventricular ring electrode 30. The use of near-field and far-field EGM sensing of arrhythmia attacks is described in U.S. Patent No. 5,193,535 (Bardy). Annotations of sensed events that can be displayed and stored with EGM data are described in US Pat. No. 4,374,382 (Markowitz).

[0053] Figure 2Only two sensing channels may be suggested in the ICD 10, an atrial sensing channel including amplifier 204 and a ventricular sensing channel including amplifier 200, however, it should be appreciated that the techniques disclosed herein may be applied to one or more cardiac electrical signals acquired using any combination of available electrodes. In some examples, a first cardiac electrical signal is acquired between the ICD housing 15 and the RV coil electrode 24, a second cardiac electrical signal is acquired between the RV coil electrode 24 and the SV coil electrode 26, and a third cardiac electrical signal is acquired between the RV tip electrode 28 and the RV ring electrode 30. The processor 224 may collect and use all three signals for analyzing the R wave and RRI and detecting atrial arrhythmias and / or ventricular arrhythmias. As described below in conjunction with Fig.11 As discussed, when a tachyarrhythmia episode is detected, at least two cardiac signals may be stored in the memory 226 for later transmission by the telemetry circuit 330. When an atrial tachyarrhythmia is detected, with or without a concurrent detection of a ventricular tachyarrhythmia, two signals having two different gain settings may be stored to provide two different signals for display on the external device 40. One signal displayed at a higher gain may result in R-wave clipping, but the relatively small amplitude P wave may be more easily observed, which enables the clinician to observe any relationship between the detected atrial tachyarrhythmia and the ventricular tachyarrhythmia (if present) by comparing the two different signals. When a ventricular tachyarrhythmia is detected without an atrial tachyarrhythmia detection, two signals may be stored, both having a gain setting that avoids R-wave clipping.

[0054] Telemetry circuitry 330 includes a transceiver for receiving downlink telemetry from external device 40 and transmitting uplink telemetry to external device 40 using antenna 332. Telemetry circuitry 330 provides two-way telemetry communications with external device 40, as described above.

[0055] The ICD 10 may receive programmable operating parameters and algorithms via the telemetry circuit 330 for storage in the memory 226 and accessed by the processor 224 for controlling ICD functions. For example, instead of RAM 226 or in addition to RAM 226, the ICD 10 may include any volatile, non-volatile, magnetic, optical, or electrical medium, such as random access memory (RAM), read-only memory (ROM), non-volatile RAM (NVRAM), electrically erasable programmable ROM (EEPROM), flash memory, or any other digital medium. The memory 226 may be accessed by the processor 224 for controlling ICD functions. For example, the cardiac rhythm detection parameters and treatment control parameters used by the ICD 10 may be programmed via the telemetry circuit 330. Thus, the memory 226 of the IMD 10 may store program instructions, which may include one or more program modules executable by the processor 224. When executed by the processor 224, such program instructions may cause the processor 224 and the IMD 10 to provide the functions attributed to them herein. Program instructions may be embodied in software, firmware, and / or RAM devices.

[0056] Data stored or acquired by ICD 10, including physiological signals or associated data derived therefrom, device diagnostics, and a history of detected arrhythmia episodes and delivered therapies, may be retrieved from ICD 10 by external device 40 pursuant to an interrogation command received by telemetry circuitry 330. Data to be uplinked to the external device and control signals for telemetry circuitry 330 are provided by processor 224 via address / data bus 218. Received telemetry is provided to processor 224 via multiplexer 220. Numerous types of telemetry systems known for implantable medical devices may be implemented in ICD 10.

[0057] exist Figure 2 Other circuit systems shown in illustrate therapy delivery circuit systems that may be included in an ICD or other implantable medical device that employs the atrial arrhythmia detection techniques disclosed herein when the device is configured to provide cardiac pacing, cardioversion, and defibrillation therapy. For example, the pacemaker timing and control circuit system 212 may include a programmable digital counter that controls basic time intervals associated with various single-chamber, dual-chamber, or multi-chamber pacing modes or anti-tachycardia pacing therapies delivered in the atria or ventricles. The pacemaker timing and control circuit system 212 also sets the amplitude, pulse width, polarity, or other characteristics of cardiac pacing pulses under the control of the processor 224.

[0058] During pacing, an escape interval counter within the pacemaker timing and control circuitry 212 is reset when an R wave and a P wave are sensed as indicated by signals on lines 202 and 206, respectively. In accordance with the selected pacing mode, pacing pulses are generated by atrial pacing output circuit 214 and ventricular pacing output circuit 216. Pacing output circuits 214 and 216 are coupled to the desired electrodes for pacing via switch matrix 208. The escape interval counter is reset when a pacing pulse is generated and thereby controls the basic timing of cardiac pacing functions, including anti-tachycardia pacing.

[0059] The duration of the escape interval is determined by the processor 224 via the data / address bus 218. When the escape interval counter is reset by a sensed R wave or P wave, the count value present in the escape interval counter can be used to measure the RR interval and PP interval for detecting the occurrence of various arrhythmias. The processor 224 can also track the number of pacing pulses delivered, particularly the number of ventricular pacing pulses delivered, over a predetermined time period as a factor for classifying the cardiac electrical signals during the time period.

[0060] Processor 224 includes an associated read-only memory (ROM) in which resides stored programs that control the operation of processor 224. A portion of memory 226 may be configured as a plurality of recirculating buffers capable of storing a series of measured intervals for analysis by processor 224 for predicting or diagnosing arrhythmias.

[0061] In response to detecting a tachycardia, anti-tachycardia pacing therapy may be delivered by loading a protocol from processor 224 into pacemaker timing and control circuitry 212 based on the type of tachycardia detected. In the event that a higher voltage cardioversion or defibrillation pulse is required, processor 224 activates cardioversion and defibrillation control circuitry 230 to initiate charging of high voltage capacitors 246 and 248 via charging circuit 236 under control of high voltage charge control line 240. The voltage on the high voltage capacitors is monitored via voltage capacitor (VCAP) line 244, which passes through multiplexer 220. When the voltage reaches a predetermined value set by processor 224, a logic signal is generated on capacitor full (CF) line 254, thereby terminating charging. Defibrillation and cardioversion pulses are delivered to the heart by output circuit 234 via control bus 238 under control of pacemaker timing and control circuitry 212. Output circuit 234 determines the electrodes used to deliver the cardioversion or defibrillation pulses and the pulse waveform.

[0062] When the ICD 10 is coupled to Figure 1AWhen the RA lead 11 shown in is connected, atrial electrical stimulation therapy can be delivered in response to detecting an atrial tachyarrhythmia using the techniques disclosed herein. In some examples, atrial pacing and / or atrial cardioversion / defibrillation shocks can be delivered to terminate a persistent atrial tachyarrhythmia.

[0063] In some examples, the ICD 10 can be equipped with a patient notification system 250. Any patient notification method known for implantable medical devices can be used, such as generating a perceptible twitch stimulus or an audible sound. The patient notification system can include an audio transducer that emits an audible sound, including an audible statement or a musical tone stored in an analog memory and associated with a programming or query operating algorithm or with a warning trigger event as generally described in U.S. Patent No. 6,067,473 (Greeninger et al.). In some examples, the ICD 10 provides a response to atrial tachyarrhythmia detection by generating a patient notification via the system 250 and / or a clinician notification using the telemetry circuit 330. The atrial tachyarrhythmia response provided by the ICD 10 can include determining the AF burden as the total combined duration of all detected AF episodes during a predetermined monitoring time interval (e.g., 24 hours), and generating a patient notification and / or a clinician notification when the AF burden exceeds a threshold.

[0064] Figure 3A and Figure 3B is a conceptual diagram of an alternative ICD system 100 that may be configured to detect AF according to the techniques disclosed herein. Figure 3A is a front view of extracardiovascular ICD system 100 implanted in patient 112 . Figure 3B is a side view of an ICD system 100 implanted in a patient 112. ICD system 100 includes ICD 110 connected to extracardiac electrical stimulation and sensing leads 116. ICD system 100 may further include an intracardiac pacemaker 101 configured to deliver pacing pulses to a ventricular or atrial chamber.

[0065] ICD 110 includes a housing 115 that forms a hermetic seal to protect the internal components of ICD 110. The internal device components may include Figure 2 , such as, sense amplifier(s), A / D converter, pacing output circuitry, high voltage output circuitry, and processor, as well as memory and / or other control circuitry. Housing 115 of ICD 110 may be formed of a conductive material such as titanium or a titanium alloy. Housing 115 may serve as a housing electrode (sometimes referred to as a can electrode). In the examples described herein, housing 115 may serve as a means for delivering the charge supplied by HV charging circuit 236 ( Figure 2) delivered by the active can electrode of the cardioversion / defibrillation (CV / DF) shock or other high voltage pulse. In other examples, the housing 115 can be used to sense cardiac signals or for delivering unipolar low voltage cardiac pacing pulses by the pacemaker output circuit in conjunction with the lead-based cathode electrode. In other instances, the housing 115 of the ICD 110 can include multiple electrodes on an external portion of the housing. The external portion(s) of the housing 115 that serve as the electrode(s) can be coated with a material, such as titanium nitride.

[0066] ICD 110 includes a connector assembly 117 (also referred to as a connector block or header) that includes an electrical feedthrough through housing 115 to provide electrical connection between conductors extending within lead bodies 118 of leads 116 and electronic components included within housing 115 of ICD 110. Figure 2 As described, housing 115 may house one or more processors, memory, a telemetry transceiver, sensing circuit systems (such as, sensing amplifiers and analog-to-digital converters), therapy delivery circuit systems (such as, pacemaker timing and control, CV / DF control, pacing output and HV output circuits and associated charging circuits), a switch matrix, a data bus, one or more batteries or other power sources, and other components for sensing cardiac electrical signals, detecting cardiac rhythms, and controlling and delivering electrical stimulation pulses to treat abnormal cardiac rhythms.

[0067] Lead 116 includes an elongated lead body 118 having a proximal end 127 including a lead connector (not shown) configured to connect to the ICD connector assembly 117 and a distal portion 125 including one or more electrodes. Figure 3A and Figure 3B , distal portion 125 of lead 116 includes defibrillation electrodes 124 and 126 and pacing / sensing electrodes 128, 130, and 131. In some cases, defibrillation electrodes 124 and 126 may form a defibrillation electrode together because they may be configured to be activated simultaneously. Alternatively, defibrillation electrodes 124 and 126 may form separate defibrillation electrodes, in which case each of electrodes 124 and 126 may be independently activated. In some instances, defibrillation electrodes 124 and 126 are coupled to electrically isolated conductors, and ICD 110 may include a switching mechanism to allow electrodes 124 and 126 to be used as a single defibrillation electrode (e.g., activated simultaneously to form a common cathode or anode) or as separate defibrillation electrodes (e.g., activated individually, one as a cathode and one as an anode; or activated one at a time, one as an anode or cathode and the other remains inactive, with housing 115 as the active electrode).

[0068] Electrodes 124 and 126 (and in some examples, housing 115) are referred to herein as defibrillation electrodes because they are used, alone or together, to deliver high voltage stimulation therapy (e.g., cardioversion or defibrillation shock). Electrodes 124 and 126 can be elongated coil electrodes and typically have a relatively high surface area for delivering high voltage electrical stimulation pulses compared to low voltage pacing and sensing electrodes 28, 30, and 31. However, electrodes 124 and 126 and housing 115 can also be used to provide pacing functions, sensing functions, or both pacing and sensing functions in addition to or in lieu of high voltage stimulation therapy. In this sense, the use of the term "defibrillation electrodes" herein should not be construed as limiting electrodes 124 and 126 to only high voltage cardioversion / defibrillation shock therapy applications. Electrodes 124 and 126 can be used in a pacing electrode vector to deliver extracardiovascular pacing pulses (such as, ATP pulses), post-shock pacing, or other pacing therapies and / or in a sensing vector to sense cardiac electrical signals for detecting atrial and ventricular arrhythmias, commonly referred to as "cardiac events," including AF, VT, and VF.

[0069] Electrodes 128, 130, and 131 are relatively small surface area electrodes for delivering low voltage pacing pulses and for sensing cardiac electrical signals. Electrodes 128, 130, and 131 are referred to as pacing / sensing electrodes because they are generally configured for low voltage applications, e.g., as cathodes or anodes for delivering pacing pulses and / or sensing cardiac electrical signals. Electrodes 124, 126, 128, 130, and / or 131 can be used to acquire cardiac electrical signals for AF detection according to the techniques disclosed herein.

[0070] Lead 116 extends subcutaneously or submuscularly from connector assembly 127 of ICD 110 centrally over thorax 132 toward the center of the torso of patient 112 (e.g., toward xiphoid process 120 of patient 112). At a location near xiphoid process 120, lead 116 bends or turns upward and extends in a substernal position within anterior mediastinum 136. Lead 116 of system 100 is at least partially implanted beneath sternum 122 of patient 112.

[0071] The anterior mediastinum 136 can be viewed as being bounded laterally by the pleura, posteriorly by the pericardium 138, and anteriorly by the sternum 122. In some instances, the anterior wall of the anterior mediastinum 136 can also be formed by the transverse thoracic muscle and one or more costal cartilages. The anterior mediastinum 136 includes a certain amount of loose connective tissue (such as, cellulite), adipose tissue, some lymphatic vessels, lymph glands, substernal musculature, small side branches of the internal thoracic artery or vein, and the thymus. In one example, the distal portion 125 of the lead 116 extends substantially along the posterior side of the sternum 122 within the loose connective tissue and / or substernal musculature of the anterior mediastinum 136.

[0072] Leads that are implanted so that the distal portion 125 is substantially within the anterior mediastinum 136 may be referred to as "substernal leads." Figure 3A and Figure 3B , the lead 116 extends substantially centrally under the sternum 122. However, in other examples, the lead 116 may be implanted such that it extends in a position laterally offset from the center of the sternum 122. In some examples, the lead 116 may extend laterally such that the distal portion 125 of the lead 116 is below / under the thoracic cavity 132 (in addition to or in lieu of the sternum 122). In other examples, the distal portion 125 of the lead 116 may be implanted in other extracardiovascular intrathoracic locations (including the pleural cavity), or around and near the perimeter of the pericardium 138 of the heart 102 but typically not within it.

[0073] In other examples, lead 116 may remain outside the thorax and extend subcutaneously or submuscularly over thorax 132 and / or sternum 122. The path of lead 116 may depend on the location of ICD 110, the placement and location of electrodes carried by lead distal portion 125, and / or other factors.

[0074] Electrical conductors (not shown) extend from the lead connector at the proximal lead end 127 through one or more lumens of the elongated lead body 118 of the lead 116 to electrodes 124, 126, 128, 130, and 131 positioned along the distal portion 125 of the lead body 118. The lead body 118 of the lead 116 may be formed of a non-conductive material (including silicone, polyurethane, fluoropolymers, mixtures thereof, and other suitable materials) and is shaped to form one or more lumens within which one or more conductors extend. However, the technology disclosed herein is not limited to such a configuration or to any particular lead body design.

[0075] The elongated electrical conductors contained within lead body 118 are each electrically coupled to respective defibrillation electrodes 124 and 126 and pacing / sensing electrodes 128, 130, and 131. Each of pacing and sensing electrodes 128, 130, and 131 is coupled to a respective electrical conductor, which may be a separate respective conductor within the lead body. The respective conductors electrically couple electrodes 124, 126, 128, 130, and 131 to circuitry, such as a switch matrix or other switching circuitry for selection and coupling to a sense amplifier or other cardiac event detection circuitry and / or to a therapy output circuit (e.g., a pacing output circuit or HV output circuit for delivering CV / DF shock pulses). Connections between the electrode conductors and the ICD circuitry are made via connections in connector assembly 117, including associated electrical feedthroughs through housing 115. The conductors transmit therapy from the output circuitry within the ICD 110 to one or more of the defibrillation electrodes 124 and 126 and / or the pacing / sensing electrodes 128, 130, and 131, and transmit sensed electrical signals from the defibrillation electrodes 124 and 126 and / or the pacing / sensing electrodes 128, 130, and 131 to the sensing circuitry within the ICD 110.

[0076] ICD 110 may obtain electrical signals corresponding to electrical activity of heart 102 via a combination of sensing vectors including a combination of electrodes 128, 130, and / or 131. In some examples, housing 115 of ICD 110 is used in combination with one or more of electrodes 128, 130, and / or 131 in the sensing electrode vector.

[0077] ICD 110 may even obtain cardiac electrical signals using sensing vectors that include one or both defibrillation electrodes 124 and / or 126 (e.g., between electrodes 124 and 126) or a sensing vector that combines one of electrodes 124 or 126 with one or more of electrodes 128, 130, 131 and / or housing 115.

[0078] The ICD 110 analyzes cardiac electrical signals received from one or more of the sensing vectors to monitor abnormal heart rhythms, such as AF, VT, and VF. The ICD 110 generates and delivers electrical stimulation therapy in response to detection of a ventricular tachyarrhythmia (e.g., VT or VF). The ICD 110 can deliver ATP in response to VT detection, and in some cases, can deliver ATP before a CV / DF shock or during charging of a high-voltage capacitor in an attempt to avoid the need to deliver a CV / DF shock. The ICD 110 can deliver a CV / DF shock pulse when VF is detected or when VT is not terminated by the ATP.

[0079] In other examples, lead 16 may include less than three pacing / sensing electrodes or more than three pacing / sensing electrodes and / or a single defibrillation electrode or more than two electrically isolated or electrically coupled defibrillation electrodes or electrode segments. Pacing / sensing electrodes 28, 30, and / or 31 may be located elsewhere along the length of lead 16. For example, lead 16 may include a single pacing / sensing electrode 30 between defibrillation electrodes 24 and 26 and no pacing / sensing electrodes distal to defibrillation electrode 26 or proximal to defibrillation electrode 24. Various example configurations of cardiovascular external leads and electrodes and sizes that may be implemented in conjunction with the AF detection techniques disclosed herein are described in commonly assigned U.S. Patent Application No. 14 / 519,436, U.S. Patent Application No. 14 / 695,255, and provisionally filed U.S. Patent Application No. 62 / 089,417.

[0080] ICD 110 is shown as being implanted subcutaneously on the left side of patient 112 along thorax 132. In some instances, ICD 110 may be implanted between the left posterior axillary line and the left anterior axillary line of patient 112. However, ICD 110 may be implanted at other subcutaneous or submuscular locations within patient 112. For example, ICD 110 may be implanted in a subcutaneous pocket in the pectoralis region. In this case, lead 116 may extend from ICD 110 subcutaneously or submuscularly toward the manubrium of sternum 122, and bend or turn downward from the manubrium subcutaneously or submuscularly and extend to a desired location. In yet another example, ICD 110 may be placed in the abdomen.

[0081] In some patients, an intracardiac pacemaker 101 may be present in the right ventricle, in the right atrium, or along the left ventricle. The pacemaker 101 may be configured to deliver pacing pulses in response to detection of VT or according to other pacing therapy algorithms in the absence of a sensed intrinsic heartbeat. For example, the pacemaker 101 may be implanted in the right ventricle of a patient for providing single-chamber ventricular pacing. The techniques disclosed herein for classifying cardiac signals may be utilized in the presence of ventricular pacing delivered by an ICD 110 and / or by an intracardiac pacemaker such as the pacemaker 101. The pacemaker 101 may correspond generally to the intracardiac pacemaker disclosed in U.S. Pat. No. 8,923,963 (Bonner et al.). The ICD 110 may be configured to detect pacing pulses delivered by the pacemaker 101. The frequency of the pacing pulses delivered by the pacemaker 101 may be a factor determined when classifying cardiac electrical signal time periods for AF detection purposes.

[0082] The pacemaker 101 may have limited processing power and therapy delivery capabilities compared to the ICD 110, such that the advanced rhythm detection techniques disclosed herein may be implemented in the ICD 110 rather than in the pacemaker 101. Thus, the methods disclosed herein are described in conjunction with the ICD 10, ICD 10' or ICD 110 or the cardiac monitoring device 60. However, these techniques are not to be considered limited to implementation in an ICD or a subcutaneous or external cardiac monitor. Aspects of the AF detection techniques disclosed herein may be implemented in whole or in part in the pacemaker 101.

[0083] Figure 4 is a schematic diagram of a method for detecting a cardiac event by a medical device such as an ICD 10, ICD 10', cardiac monitoring device 60 or ICD 110 according to an example. Single chamber devices have been designed to detect AF using ventricular EGM signals. Illustrative methods and devices for detecting AF using ventricular EGM signals are generally described in commonly assigned U.S. Patent Application Nos. 14 / 520,798, 14 / 520,938 and 14 / 520,847 (Cao et al.). R waves accompanying ventricular depolarization are sensed from the ventricular EGM signal and used to determine the RRI, i.e., the interval between consecutive R waves. The difference in consecutive RRIs is determined by subtracting the RRI from the immediately preceding RRI. Analysis of Lorentz plots of the difference in consecutive RRIs may reveal the RRI variability pattern typical of AF.

[0084] Ritscher et al., in U.S. Pat. No. 7,031,765, generally disclose a method based on plotting a Lorentz scatter plot (such as, Figure 4 Methods for detecting atrial arrhythmias by determining irregularities in the ventricular cycle as determined by differences in the RRIs that exhibit distinguishing characteristics as shown in the Figures 1 and 2. Other methods are generally disclosed by Sarkar et al. in U.S. Pat. No. 7,623,911 and in U.S. Pat. No. 7,537,569 and by Houben in U.S. Pat. No. 7,627,368.

[0085] In the following description, reference is made to Figure 2 Circuit system and Figure 1A AF detection techniques are described in the context of an ICD 10. However, it should be understood that the methods and techniques described below may be used in Figure 1B ICD10', Figure 3A and Figure 3B ICD 110 or such Figure 1C To perform these AF detection techniques, all of the aforementioned devices may include a processor, a memory, and a sensing circuit system, such as a combination of Figure 2 Overall description.

[0086] To determine whether AF has occurred, the processor 224 ( Figure 2 ) can be based on the sensed R-wave (e.g., Figure 2 The processor 224 may determine the difference between the RRIs based on the ROUT signal line 202 in the RRI. The processor 224 may make a determination as to whether an AF event has occurred based at least in part on the pattern or characteristics of the resulting RRI differences. As described below, when the resulting characteristics of the RRI differences acquired within a predetermined time period indicate that AF is occurring, the cardiac signal time period is classified as AF. When a required number of time periods are classified as AF, AF is detected. The techniques disclosed herein may be used as part of an overall rapid arrhythmia detection and discrimination algorithm implemented in the ICD 10 or other devices described above, or in other implantable or external cardiac devices (such as, intracardiac pacemakers, leadless pacemakers, or external devices).

[0087] By generating Figure 4 The Lorenz scatter plot shown in 150 demonstrates the concept of using the characteristics of RRI differences for detecting AF. Processor 224 determines the differences (δRR) between pairs of consecutive RR intervals that can be plotted against the time series of RRI. The Lorenz plot 150 is composed of δRR along the x-axis 152. i and δRR along the y-axis 154 i-1 Thus, each plotted point in the Lorentz diagram is represented by a coordinate system equal to δRR i The x-coordinate and δRR i-1 The y-coordinate of δRR is defined as i is the difference between the ith RRI and the previous RRI (RRI i-1 ) between δRR i-1 It is RRI i-1 Compared with the previous RRI (RRI i-2 ) between the two.

[0088] Thus, each data point plotted on the Lorenz diagram 150 represents an RRI pattern involving three consecutive RRIs: the RRI measured between four consecutive sensed R-waves i , RRI i-1 , and RRI i-2 The RRI information is not limited to the detection of R waves and the determination of RRI. As used herein, the terms RRI and δRR iGenerally refers to the measurement of ventricular cycle length (VCL) and the difference between two consecutive VCL measurements, respectively, whether the VCL measurement is derived from a series of sensed R waves from the cardiac electrical signal or from a series of ventricular cycle event detections made from another physiological signal (e.g., peak pressure determined from a pressure signal). For the purpose of illustration, the methods described herein refer to R wave detection and the difference between (δRR) for performing VCL measurements. i , δRR i-1 ) point determination.

[0089] As in Figure 4 As shown in FIG. 1 , a series of R waves 170 (represented by vertical bars) are sensed, and to plot points on the Lorentz plot region 150, the RRIs are determined by determining the successive RRIs determined from the sensed R waves 170 (δRR i , δRR i-1 ) point. In the example shown, three consecutive RRI (RRI i-2 , RRI i-1 and RRI i ) provides the first data point 155 on the Lorenz plot region 150. i-2 With RRI i-1 The difference between the δRR i-1 Close to 0. δRR i , i.e. RRI i-1 and RRI i The difference between is a positive change. Correspondingly, in the Lorentz diagram 150, (δRR) with a y-coordinate close to 0 and a positive x-coordinate is plotted i , δRR i-1 ) point 155, representing a first series 172 of four sensed R-waves (three RRIs).

[0090] The next series 174 of three RRIs provides the next (delta RRI) with a negative x-coordinate (the last RRI of the series 174 is less than the immediately preceding RRI) and a positive y-coordinate (the middle RRI of the series 174 is longer than the first RRI of the series). i , δRR i-1 ) point 156. This plot (δRR i , δRR i-1 ) point process continues to provide a three-period series 176 of data points 158, etc.

[0091] Figure 5 is a diagram of a two-dimensional histogram representing a region 150 of a Lorenz plot for detecting atrial tachyarrhythmias used in the techniques disclosed herein. Figure 4The Lorentz diagram region 150 shown in FIG. 1 is digitally represented by a two-dimensional histogram 180, which is respectively i Coordinate (corresponding to the x-axis) and δRR i-1 The coordinate (corresponding to the y-axis) has predefined ranges 184 and 186 in both the positive and negative directions. The two-dimensional histogram 180 is divided into two regions each having a delta RR i and δRR i-1 In one example, for δRR i and δRR i-1 For both values, the histogram range can extend from -1200 ms to +1200 ms, and the histogram range can be divided into bins extending up to a range of 7.5 ms in each of the two dimensions, resulting in a 160 bin x 160 bin histogram 180. The consecutive RRI differences determined over the classification period are used to populate the histogram 180. Each bin stores the (δRR) that falls within each corresponding bin range. i , δRR i-1 ) count of the number of data points. The bin counts may then be used by processor 224 to determine RRI variability metrics and patterns for use in detecting rhythm types.

[0092] The RRI variability measure is determined from the histogram bin counts. In general, the more occupied histogram bins are, the smaller the (δRR i , δRR i-1 ) points, the more irregular the VCL during the data acquisition period. Thus, one measure of RRI variability that can be used to detect AF (which is associated with a highly irregular VCL) can take into account the number of histogram bins that have a count of at least one (which are referred to as "occupied" bins). In one example, a measure of RRI variability for detecting AF (referred to as an AF score) is determined by processor 224 as generally described in the '911 patent referenced above. Briefly, the AF score can be defined by the following formula:

[0093] AF score = irregularity evidence - origin count - PAC evidence

[0094] Wherein, the evidence of irregularity is the number of occupied histogram bins outside the Zero Segment 188 defined around the origin of the Lorentz plot region. During normal sinus rhythm or highly organized atrial tachycardia, almost all points will fall into the Zero Segment 188 due to the relatively small constant difference between consecutive RRIs. Therefore, a large number of occupied histogram bins outside the Zero Segment 188 is positive evidence of AF.

[0095] The origin count is the number of points defined around the origin of the Lorenz diagram in the zero segment 188. A high origin count indicates a regular RRI (a negative indicator of AF) and is therefore subtracted from the irregularity evidence term. In addition, a regular PAC evidence score can be calculated as generally described in the '911 patent cited above. The regular PAC evidence score is calculated based on a cluster feature pattern of data points, which is particularly associated with atrial premature beats (PACs) (PACs occur with regular coupling intervals and present a regular RRI pattern), for example, associated with bigeminy (short short long RRI) or tripeminy (short short long RRI). In other embodiments, as described in any of the '765, '316, '911, '569 and '368 patents cited above, an AF score and / or other RRI variability score for classifying atrial rhythms may be determined by the processor 224. Methods for rejecting noise in determining points of a Lorentz plot and AF scores are generally disclosed in U.S. Pat. No. 8,639,316 (Sarkar et al.). Methods for adjusting AF scores based on the presence of ectopic sites can be used in the techniques disclosed herein and are generally disclosed in U.S. Pat. No. 8,977,350 (Sarkar et al.). Other techniques that can be used to calculate AF scores are generally disclosed in U.S. Patent Application Nos. 14 / 695,135, 14 / 695,156, 14 / 695,171, and 14 / 695,111 (Sarkar et al.), all filed on April 24, 2015.

[0096] The AF score is compared to an AF score threshold for classifying a predetermined time period of the cardiac signal as AF or non-AF based on the RRI analysis. The AF score threshold can be selected and optimized based on historical clinical data of a selected patient population or individual patient historical data, and the optimal AF score threshold setting may vary from patient to patient. In an illustrative example, the AF score can have a possible range of 0 to 100. The AF score threshold can be set between 25 and 75. If the AF score meets the AF score threshold or crosses the AF score threshold, the time period (during which the RRI is collected and therefore the cardiac signal occurs during this time period) is classified as an AF time period. The AF score threshold can be adjusted after at least one time period in the cardiac signal is classified as AF, and the adjusted AF score threshold can be used to classify subsequent time periods that may result in AF detection. In the example given above where the maximum AF score is 100 and the initial AF score threshold is at least 26 and not more than 75, the adjusted AF score threshold is less than the initial AF score threshold and can have a value ranging from 19 to 57. Thus, the adjusted AF score threshold may be between 65-85% of the initial AF score threshold, and in some instances, between 70-75% of the initial AF score.

[0097] An AF detection is made when a threshold number of time periods are classified as AF. In one example, a single n-second or n-minute time period classified as AF based on an AF score that meets an AF score threshold may result in an AF detection. In other examples, a higher number of time periods may need to be classified as AF before the heart rhythm is detected as AF.

[0098] Processor 224 provides a response to AF detection, which may include stopping, adjusting, or delivering therapy (e.g., stopping ATP or shock therapy used to treat ventricular tachyarrhythmias, or delivering ventricular anti-tachyarrhythmia therapy (if available)), storing cardiac signal data that may later be retrieved by a clinician using an external device 40, triggering a patient notification system 250, transmitting data via telemetry circuitry 330 to alert a clinician, and / or triggering other signal acquisition or analysis.

[0099] After AF detection is performed, RRI analysis may continue to be performed by processor 224 to populate the histogram during the next n-second detection period. After each detection period, the AF score may be re-determined and the histogram bins reinitialized to zero for the next detection period. The new AF score (or other RRI variability measure) determined at the end of each detection period may be used to determine whether the AF episode is maintained or terminated after the initial AF detection is performed.

[0100] Figure 6 is a flowchart 300 of a method for determining factors for classifying time periods for detecting atrial arrhythmias according to one example. Flowchart 300, as well as other flowcharts presented herein, are intended to illustrate the functional operation of an ICD 10 or another device that performs the disclosed method and should not be construed as reflecting the specific form of software, firmware, or hardware required to practice the method. It is believed that the specific form of the software will be determined primarily by the specific system architecture employed in the device and the specific detection and therapy delivery methods employed by the device. It is within the capabilities of a person skilled in the art to provide software, firmware, and / or hardware to implement the techniques disclosed herein in the context of any modern medical device, given the disclosure herein.

[0101] The methods described in conjunction with the flowcharts presented herein may be implemented in a non-transitory computer-readable medium including instructions for causing a programmable processor (such as processor 224) to implement the described methods. "Computer-readable medium" includes, but is not limited to, any volatile or non-volatile medium, such as RAM, ROM, CD-ROM, NVRAM, EEPROM, flash memory, etc. The instructions may be implemented as one or more software modules, which may be executed by themselves or in combination with other software.

[0102] like Figure 6 As shown, the processor 224 identifies a ventricular event (such as, based on the R wave of the Rout signal line 202) at box 301, and identifies the ventricular event as an intrinsic sensed event Vs or a paced event Vp caused by pacing delivered by the ICD 10 or ICD 10' (or by the ICD 110 or the pacemaker 101). Depending on the number of RR intervals selected for determining the RR interval difference, the processor 224 determines whether a predetermined number of events (ventricular pacing events Vp or intrinsic ventricular sensed events VS) have been identified at box 302. For example, according to one example, if the desired number of RR intervals for determining consecutive RR interval differences is three, the predetermined number of events used in box 302 will be four events, where the four events form a sensing window. If the predetermined number of events has not yet been reached (the "No" branch of box 302), the processor 224 determines the next ventricular event at box 301 and repeats the process.

[0103] Once a predetermined number of events are identified (the "yes" branch of block 302), an event window is identified based on the four events at block 304, and a determination can be made at block 306 as to whether the number of events in the event window that are ventricular pacing Vp events is less than or equal to a predetermined pacing event threshold. For example, according to one example, the pacing event threshold is set to 1, causing processor 224 to determine whether one or less of the identified events in the event window are ventricular pacing events. If the number of identified events in the event window that are ventricular pacing Vp events is not less than or equal to (i.e., greater than) the predetermined pacing event threshold (the "no" branch of block 306), processor 224 identifies the next event at block 301, and the process repeats.

[0104] If the number of events in the event window that are ventricular pacing Vp events is less than or equal to the predetermined pacing event threshold (the "yes" branch of box 306), the processor 224 determines whether each of the RR intervals associated with the events in the current event window is greater than the predetermined interval threshold at box 308. For example, according to one example, the processor 224 determines whether each of the RR intervals associated with the events in the event window is greater than 220 milliseconds. If each of the RR intervals associated with the events in the event window is not greater than the predetermined interval threshold (the "no" branch of box 308), the processor 224 identifies the next event at box 301 and repeats the process using the next identified event and the resulting next event window.

[0105] If each of the RRIs associated with the events in the event window is greater than a predetermined interval threshold (the "yes" branch of box 308), the processor 224 determines the difference between consecutive RRIs associated with the identified events in the event window at box 310. Once the RRI difference for the current event window has been determined at box 308 to populate the Lorenz plot histogram as described above, the processor 224 determines at box 312 whether a predetermined time period has expired. The processor 224 can set a timer or counter to control the acquisition of RRI differences within a predetermined time period at the beginning of the method of flowchart 300. In one example, the predetermined time period can be set to two minutes. In other examples, the predetermined time period can be between one and five minutes. If the time period has not expired (the "no" branch of box 312), the processor 224 returns to box 301 to identify the next ventricular event and repeats the process using the next event and the resulting next event window.

[0106] Once the timer expires (the "yes" branch of block 312), the processor 224 determines an AF score based on the RRI differences determined during a predetermined time period (e.g., two minutes) at block 314. The AF score may be determined as described above with reference to Figure 5 and / or as described in the cited patents. Figure 7 As described, the determined AF score for the predetermined time period is used to classify the time period (and therefore the cardiac signal during the time period) as an AF time period, a non-AF time period, or an unclassified time period. The stored RRI difference is then cleared and all counters and timers are reset at box 316. A timer set to a predetermined time period (e.g., two minutes) is reset. At box 300, the processor 224 identifies the next ventricular event and repeats the process for the next time period using the next identified event and the next event window.

[0107] Figure 74 is a flow chart of a method for classifying a predetermined time period according to an example. The predetermined time period will be described in the context of a predetermined time period having a length of two minutes. Figure 7 The example described in the flowchart 400. However, Figure 7 The techniques described in or elsewhere throughout this specification may be for predetermined time periods that are longer or shorter than two minutes. Once the predetermined time period (e.g., a two-minute time period) has expired and the Lorenz diagram has been populated with points associated with each determined RR interval difference determined based on the intervals in each event window that occurred during the two-minute time period (e.g., in conjunction with Figure 4 , Figure 5 and Figure 6 As described above, the processor 224 determines whether to classify the time period as an AF time period, a non-AF time period, or an unclassified time period (i.e., the time period may be neither classified as an AF time period nor classified as a non-AF time period). For example, the processor may analyze one or more of the factors in any combination or specific order to make the determination.

[0108] As combined Figure 7 As described in the example of , the factors that may be analyzed for classifying the two minute time periods are the number of valid RRI difference pairs, RRI length, number of paced beats, number of short intervals, presence of ventricular event oversensing, presence of T wave oversensing, detection of ventricular tachyarrhythmias (e.g., SVT, VT, or VF or more generally referred to as "other episodes"), and AF score. However, processor 224 may analyze only these factors and / or a subset including other factors.

[0109] The processor 224 may determine at box 401 the number of RRI difference pairs acquired during a two-minute time period, each RRI difference pair representing a point of the Lorenz diagram. A determination is then made at box 402 as to whether the total number of RRI difference pairs formed during the two-minute time period is greater than an interval pair threshold. According to one example, the threshold number of RRI difference pairs is set to 30, but other thresholds may be used. If the total number of RRI difference pairs (representing three consecutive RRIs) during the two-minute time period is less than the threshold (the "yes" branch at box 402), the two-minute time period is determined to be unclassified at box 404. In the example described above where the threshold is set to 30, the "yes" branch of box 402 means that less than 30 RRI difference pairs are determined during the two-minute time period, resulting in the Lorenz diagram histogram being filled with less than 30 points. An AF score determined from less than the threshold number of RRI difference pairs may not result in a reliable AF score for the predetermined time period, and therefore the AF score is not used to classify the time period as AF or non-AF.

[0110] If the number of RRI difference pairs formed during the two-minute time period is not less than the interval pair threshold (30 in the above example) (the "no" branch of box 402), the interval pair factor for classifying the time period as AF or non-AF based on the AF score determined from the RRI difference pairs is satisfied. In other words, using the above example, 30 or more RRI difference pairs are determined during the two-minute time period, resulting in the Lorenz plot histogram being populated with 30 or more points. The number of RRI difference pairs obtained during the predetermined time period is sufficient to reliably classify the time period as AF or non-AF based on the AF score. According to another example, the processor 224 may additionally or alternatively determine at box 410 the total number of RRIs determined to be less than the interval threshold applied at box 308 of flowchart 300 during the predetermined time period. If the number of RRIs exceeding the threshold (e.g., exceeding a predetermined number of RRIs or a predetermined percentage of the total number of RRIs occurring during the two-minute time period) is less than the interval threshold (the "yes" branch at box 412), the two-minute time period is determined as unclassified at box 404. If the number of RRIs less than the interval threshold does not reach or exceed a predetermined number (e.g., if less than 10 RRIs are less than the interval threshold during a two-minute time period), it is determined at box 412 that the RRI length factor is not satisfied (the "No" branch of box 412) for classifying the predetermined time interval as unclassified. Based at least on this factor, the classification of AF or non-AF based on the AF score is guaranteed.

[0111] To classify a two minute time period as AF or non-AF, the processor 224 may determine at box 414 a short interval count from the total number of RRIs in all event windows obtained during the two minute time period, the short interval count being, for example, less than or equal to a predetermined short interval threshold, such as 120 milliseconds or 130 milliseconds. The processor 224 determines at box 416 whether the short interval count is greater than the short interval threshold, such as, for example, 5 short intervals. Too many short intervals during the two minute time period indicate the possibility of ventricular oversensing of non-physiological signals (such as, EMI or lead noise due to lead breakage). In this case, the use of the RRI to determine the AF score and classify the time period as AF or non-AF based on the AF score may be unreliable.

[0112] If the determined short interval count is greater than the short interval count threshold (the "yes" branch at box 416), the two-minute time period is determined to be unclassified at box 404. On the other hand, if the short interval count is less than the short interval count threshold, the time period can be classified based on the AF score (the "no" branch at box 416). The short interval count factor minimizes false AF detection due to oversensing of lead noise.

[0113] Processor 224 may additionally or alternatively determine the number of events identified within all event windows that are determined to be ventricular pacing Vp events during the total two minute period at block 418. A determination is made at block 420 as to whether the determined number of ventricular pacing Vp events identified during all event windows of the two minute period is greater than a total ventricular pacing Vp event threshold. According to one example, the total ventricular pacing Vp threshold is set to 30 ventricular pacing Vp events, although other thresholds may be used.

[0114] If the number of ventricular pacing Vp events during the two-minute time period is greater than the total ventricular pacing Vp event threshold (the "yes" branch of box 420), the processor 224 classifies the two-minute time period as unclassified at box 404. Ventricular pacing pulses may include bradycardia pacing pulses and / or ATP pacing pulses and may be delivered by ICD 10, ICD 10', or ICD 110 or by another implanted device (e.g., pacemaker 110). On the other hand, if the determined number of ventricular pacing Vp events is not greater than the Vp event threshold (the "no" branch at box 420), the two-minute time period is not classified as unclassified; as long as no other factors lead to the determination that the time period is unclassified, the classification of AF or non-AF can be made based on the AF score.

[0115] The processor 224 may be configured to evaluate both the R wave and the RRI for detecting supraventricular tachycardia (AVT), VT, and VF while the AF detection algorithm described herein is being performed. The ICD 10 may be configured to deliver a therapy such as ATP in response to detecting a VT. Thus, if a ventricular tachyarrhythmia detection (e.g., SVT, VT, VF detection) is performed during or at the expiration of the current time period, as determined at box 421, the current time period is determined as unclassified at box 404. If no other episode detection is being performed, the process may proceed to box 422.

[0116] At box 422, processor 224 may additionally or alternatively determine whether oversensing of ventricular events caused by noise was detected during the two-minute time period. Oversensing detection may be performed by processor 224 using an implemented oversensing detection scheme (such as the oversensing detection method generally described in U.S. Patent No. 7,333,855 to Gunderson et al.). If the oversensing detection criteria are met during the two-minute time period or are in the process of meeting the oversensing detection criteria (the "yes" branch of box 422), the two-minute time period is determined to be unclassified at box 404. The detection of oversensing indicates that the RRI may be unreliable for determining the AF score and classifying the time period based on the AF score. If no oversensing detection is performed during the two-minute time period or is not in the process of performing oversensing detection (the "no" branch of box 422), AF or non-AF classification may be performed based on the AF score as long as other factors do not cause the time period to be classified as unclassified.

[0117] At box 424, the processor 224 may determine whether T wave oversensing occurred during the two-minute time period. The determination of T wave oversensing may be performed by the ICD 10 using an implemented T wave oversensing detection scheme, such as the T wave oversensing determination described in U.S. Pat. No. 7,831,304 to Gillberg et al. If a determination of T wave oversensing is made during the two-minute time period or is in progress (the "yes" branch of box 424), the T wave oversensing factor is satisfied as an indication that the two-minute time period is not classified. At box 404, the processor 224 classifies the two-minute time period as unclassified. If a determination of T wave oversensing is not made or is not in the process of making a determination of T wave oversensing within the two-minute time period (the "no" branch of box 424), the T wave oversensing factor is not satisfied. The processor 224 proceeds to box 426 to classify the time period based on the AF score.

[0118] In this way, the ICD 10 (or ICD 10', ICD 110, or cardiac monitoring device 60) can analyze cardiac electrical signals within a two-minute time period for one or more of the described factors, which, if met, will result in the two-minute time period being classified as "unclassified." In other words, the ICD 10, ICD 110, or other device can analyze all of the described factors or only a subset of the described factors when making this determination. In some examples, if at least one of the described factors for identifying a two-minute time period as unclassified is met, the two-minute time period is classified as unclassified at box 404. If it is determined according to predetermined criteria that none of the factors evaluated in boxes 401 to 424 are met, the time period is classified as AF or non-AF. In this way, the AF score is determined based on the populated Lorenz plot histogram as described above. If the AF score is greater than the AF threshold at box 426, the two-minute time period is classified as AF at box 406. On the other hand, if the AF score is not greater than the AF threshold (“NO” branch of block 426 ), then the two minute period is classified as non-AF at block 408 .

[0119] It should be understood that the determination of whether a time period is classified as unclassified (box 404), classified as AF (box 406), or classified as non-AF (box 408) can be made in any order or simultaneously, such that a determination of a two-minute time period as an unclassified time period can be used to override an initial determination of classifying the two-minute time period as AF or non-AF, or a determination of a two-minute time period as an unclassified time period can be made prior to determining an AF score for classification based on the AF score.

[0120] Figure 8 is a schematic diagram of atrial fibrillation detection that may be performed by a medical device according to one example. The description will be made in the context of a predetermined time period having a length of two minutes. Figure 8 However, Figure 8 The techniques described in can be used for a predetermined time period that is longer or shorter than two minutes.

[0121] like Figure 8 As shown in FIG. 1 , the processor 224 uses a combination of Figure 7 The described method classifies the cardiac signal for each two-minute period as AF, non-AF, or unclassified. The classification of the periods is used to detect AF episodes. For example, once a predetermined number of two-minute periods (such as three periods) have been classified as AF, the device detects an AF episode. Thus, if Figure 8As shown in the scenario of the timing diagram (a) of , once a predetermined number of two-minute time periods 500, 502, and 504 are classified as AF, the processor 224 detects an AF episode at time 505. The processor 224 can track the number of two-minute time periods classified as AF by updating an AF event counter each time a time period is classified as AF.

[0122] However, in the scenario illustrated in timing diagram (b), two consecutive two-minute time periods 506 and 508 are classified as AF, but the next two-minute time period 510 is determined to be unclassified, followed by the subsequent time period 512 being classified as AF. According to one example, the processor 224 can ignore the unclassified two-minute time period 510 and detect an AF episode at time 513 once the third time period 512 is classified as AF, so that an AF episode can be detected despite the occurrence of one or more intermittent unclassified two-minute time periods between the AF classified time periods.

[0123] In the timing diagram of scenario (b), the AF event counter may be incremented to one when a two minute time period 506 is identified. When a subsequent two minute time period 508 is identified, the AF event counter is incremented to two. When a subsequent two minute time period 510 is identified, the AF event counter remains at a count of two because the event is determined to be unclassified. When a subsequent two minute time period 512 is identified, the AF event counter is incremented to three, and an AF episode is detected in response to the AF event counter reaching the AF detection threshold (which is 3 in this example).

[0124] As shown in the timing diagram of scenario (c), classifying one or more time periods as non-AF results in no AF episode being detected. During the determination of whether a predetermined number of two-minute time periods are classified as AF, the processor 224 updates the AF event counter each time an AF classification is made as described above. For example, upon classification of two-minute time period 514, the AF event counter is incremented to 1, and upon classification of a subsequent two-minute time period 514, the AF event counter is incremented to 2. If two-minute time period 518 is also classified as AF, the processor 224 will detect an AF episode because three two-minute time periods classified as AF will occur, for example, as described in the timing diagram of scenario (a) above. However, because two-minute time period 518 is classified as non-AF, no AF episode detection is made. The non-AF classification of time period 518 may be evidence that the AF episode that caused the AF classification of time periods 514 and 516 was terminated or that the AF episode was a non-sustained AF episode. In response to classifying time period 518 as non-AF, the AF event counter is reset to zero. In other examples, when a time period is classified as non-AF, the AF counter may be decremented rather than immediately reset to zero. In the timing diagram of scenario (d), upon classification of the two-minute time period 522 as AF, the AF event counter is incremented to 1, and upon classification of the subsequent two-minute time period 524 as AF, the AF event counter is incremented to 2. Upon classification of the subsequent two-minute time periods 526 and 528, both of which are determined to be unclassified, the AF event counter remains unchanged at a count of two. Upon classification of the subsequent two-minute time period 530 as AF, the AF event counter is incremented to three, and an AF episode is detected at time 531.

[0125] If any of time periods 524, 526, 528, or 530 has been classified as non-AF, the AF event counter will be reset to zero and the process will repeat starting with the next classified two-minute interval. However, in addition to resetting the AF event counter in response to a two-minute time period being classified as a non-AF time period, the processor 224 may also be configured to reset the AF event counter to zero if a predetermined number of two-minute time periods are determined to be unclassified. For example, five consecutive two-minute time periods determined to be unclassified may cause the AF event counter to be reset. In other examples, more than five or less than five unclassified time periods (which may not need to be consecutive) may cause the AF counter to be decremented or reset to zero. Therefore, in the timing diagram of scenario (e), when the two-minute time period 532 is identified as AF, the AF event counter is incremented to one. When the subsequent two-minute time period 534 is identified as AF, the AF event counter is incremented to two. The AF event count remains unchanged at two upon identification of four subsequent two-minute time periods 536, 538, 540, and 542, all of which are determined to be unclassified. In the example shown, the next two-minute time period 546 is classified as AF. The AF event counter is incremented from 2 to 3, and in response to the AF event counter reaching a threshold count (three in this example), an AF episode is detected by the processor 224 at 545.

[0126] On the other hand, if the subsequent two-minute time period 546 is determined to be unclassified, then the AF event counter will be reset to zero in response to a threshold number (five in this example) of consecutive unclassified time periods. In the event that time period 546 is classified as a non-AF time period, the AF event counter will also be reset to zero. In either case, if time periods 532 and 534 represent true AF episodes, then the AF episodes have terminated or are not ongoing, as evidenced by the unclassified and / or non-AF classified time periods. The process is repeated starting with the next classified two-minute interval.

[0127] Fig. 9 FIG. 1 is a schematic diagram of a method for detecting atrial fibrillation that may be performed by ICD 10 (or ICD 10′, ICD 110, or cardiac monitoring device 60) according to another example. The method will be described in the context of a predetermined time period having a length of two minutes and an AF detection threshold being set equal to three time periods classified as AF. Fig. 9 However, the described techniques may utilize different time period durations and / or different thresholds. For example, the predetermined time period may be between one and five minutes, and the number of time periods classified as AF may be greater than or equal to one and less than or equal to five.

[0128] As described above, processor 224 uses a combination of Figure 7 The described method classifies the cardiac signal within each two-minute period as AF, non-AF, or unclassified. If the factors that cause the two-minute period to be unclassified are not met, each two-minute period is classified as AF or non-AF based on the AF score. Fig. 9 In the method of , the threshold to which the AF scores are compared for classifying time periods is not a fixed value, but is dynamically adjusted by the processor 224 in response to the classification of the two-minute time periods.

[0129] For example, once a predetermined number of time periods (e.g., one time period) are classified as AF based on a first AF score threshold, if a next predetermined number of time periods are classified as any combination of AF and / or unclassified, the AF score threshold is adjusted to a second, lower value. In other words, after an initial AF classification using a first, higher AF score threshold, the processor 224 adjusts the AF score threshold for the next predetermined number of consecutive time periods (e.g., one time period) as long as none of the next predetermined number of consecutive time periods are classified as non-AF. Fig. 9 The AF score threshold is reduced to a second, lower value at the expiration of the predetermined number of next consecutive time periods after the initial AF classification (two consecutive time periods after the initial AF classification in the example shown in FIG. 1 ). In other examples, the predetermined number of next consecutive time periods may be less than two (e.g., zero or one) or more than two. If any of the predetermined number of next consecutive time periods after the initial AF classification is classified as non-AF, the AF score threshold remains at the first, higher threshold.

[0130] Therefore, if Fig. 9 As shown in the scenario of the timing diagram (a) of FIG. 6 , based on the AF score determined for the time period 610 being greater than the first AF score threshold 602 and the analysis of other classification factors not resulting in unclassified time periods (such as Figure 7 604). In this example, the first AF score threshold 602 remains valid for at least two more consecutive time periods 612 and 614. If both time periods are also classified as AF, AF is detected at time 622 in response to the AF score exceeding the first AF score threshold 602. Additionally, the processor 224 adjusts the AF score threshold to a second, lower AF score threshold 604. The AF scores of subsequent time periods are compared with the lower threshold 604 for classification of the corresponding time periods.

[0131] The lower AF score threshold 604 can be set to a percentage of the initial AF score threshold 602, for example, approximately 75% of the initial AF score. For illustration, when the maximum possible value of the AF score is 100, the first AF score threshold can be set to 75 and adjusted to a second lower AF score of 57. In another example, the first AF score threshold is 60 and the second AF score threshold is 45. In still other examples, the first AF score threshold is 60 and the second AF score threshold is 45, the first AF score threshold is 50 and the second AF score threshold is 38, the first AF score threshold is 40 and the second AF score threshold is 30, or the first AF score threshold is 25 and the second AF score threshold is 19. The user can program the AF score thresholds based on a selection of: least sensitive, such as a first threshold of 75 and a second threshold of 57; most sensitive, such as a first threshold of 25 and a second threshold of 19, where other examples corresponding to less sensitive settings (first threshold of 60 and second threshold of 45), balanced sensitivity (first threshold of 50 and second threshold of 38), and more sensitive (first threshold of 40 and second threshold of 30) are given above. In other examples, the actual values ​​of the first threshold and the second threshold may be individually selectively programmable in any combination of a first range, such as from 25 to 75 and including 75 for the first threshold, and a second range, such as from 19 to 57 and including 57 for the second threshold, as long as the first threshold is greater than the second threshold. In other examples, the second threshold may be set to another percentage of the initial threshold, such as between 65-85%, between 70-80%, or between some other percentages.

[0132] By reducing the AF scores for subsequent time periods, AF detection sensitivity is increased at appropriate times while maintaining AF detection specificity by using a first, higher AF score threshold and applying factors that result in an unclassified time period. For example, even though the first, higher threshold 602 is not met (and there are no factors that result in an unclassified classification), the next two consecutive time periods 616 and 618 are both classified as AF based on the AF scores exceeding the second threshold 604. Even though the AF scores for these time periods 616 and 618 are each less than the first threshold 602, the detected AF episode is detected as still ongoing during the time periods 616 and 618. Because the AF score is less than the second threshold 604, the next time period 620 is classified as non-AF. In response to the non-AF classification, the processor 224 adjusts the AF score threshold from the lower threshold 604 back to the higher threshold 602 at time 624. Termination of the AF episode is detected in response to the non-AF classification. The AF episode duration 615 is the time interval from the earliest time period 610 classified as AF (resulting in AF detection at time 622) to the end of the last AF classification time period 618 (which precedes the termination detection at time 624, i.e., before the time period 620 classified as non-AF).

[0133] In the scenario illustrated in timing diagram (b), after the initial time period 630 is classified as AF based on the first AF score threshold 602, two consecutive two-minute time periods 632 and 634 are classified as unclassified. In response to the absence of non-AF classification for both time periods 632 and 634 after the initial AF classification of time period 630, the processor 224 adjusts the AF score threshold to the second lower threshold 604 at time 642. Three consecutive classifications including at least an initial AF classification and no non-AF classification result in the adjustment of the AF score threshold. Thus, in one example, the time periods 632 and 634 immediately and consecutively after the initial AF classification time period 630 may both be unclassified (as shown in this example), both classified as AF, or one classified as AF and one unclassified, resulting in the adjustment of the AF score threshold at time 642.

[0134] However, since only one time period 630 is classified as AF in the example shown, no AF detection is made at time 642 when the AF score threshold is adjusted. The next two time periods 636 and 638 are classified as AF in response to the AF score being less than the adjusted AF score threshold 604 (and the factor that would cause an unclassified classification to be made is not determined). When the AF event counter reaches a count of three at time 643, an AF detection is made. In this example, the next time period 640 is classified as non-AF. The termination of the AF episode is detected, and in response to the non-AF classification and the resulting episode termination detection, the AF score threshold is adjusted upward from the lower value 604 back to the higher value 602 at time 644. The episode duration 635 begins with the earliest time period 630 that was classified as AF and resulted in an AF detection at time 643, and extends through the most recent AF classified time period 638 prior to the termination detection at time 644. Episode duration 635 includes unclassified time periods 632 and 634 that do not result in detection of termination at time 644. Unclassified time periods 632 and 634 occur between AF classified time periods 630 and 636 and are therefore included in AF episode duration 635. Figure 7 Time periods 632 and 634 may be classified as unclassified due to any of the other factors described in . In one specific example, due to ventricular tachyarrhythmia detection ( Figure 7 421), one or both of time periods 632 and 634 may be classified as unclassified. By allowing time periods 632 and 634 to be classified as unclassified when a ventricular tachyarrhythmia is detected, the detection of AF and the determination of AF episode duration 635 are uninterrupted. The detection of AF episodes that coincide with ventricular tachyarrhythmia episodes provides important diagnostic information for the clinician to use in correctly determining the patient's cardiac rhythm state and subsequent treatment.

[0135] In scenario (b) and subsequent scenarios, time periods determined to be unclassified (e.g., time periods 632 and 634) are represented as having AF scores equal to the currently set AF score threshold. However, it should be understood that the actual AF score (if determined) may be greater than, equal to, or less than the current value of the AF score threshold, but when analysis of other factors causes the time period to be determined to be unclassified (such as in combination with the AF score threshold), the actual AF score may be greater than, equal to, or less than the current value of the AF score threshold. Figure 7 In some cases, if due to the combination of Figure 7 If the described analysis of one or more factors determines that a time period is unclassified, then a determination of an AF score for the current time period may not be made; in some examples, classifying a time period as unclassified may obviate the need to determine an AF score.

[0136] In scenario (c), after the initial AF classification time period 650, the AF score threshold is adjusted from the first threshold 602 to the second threshold 604 at time 674 based on the first higher threshold 602, followed by two consecutive time periods 652 and 654 that do not include non-AF classification. The processor 224 can increment the AF event counter in response to each AF classification and increment the unclassified event counter in response to each unclassified time period. Accordingly, in the example of scenario (c), at time 674, the AF event counter is at count one and the unclassified event counter is at count two. After time period 656, the unclassified event counter is at count three, and after time period 658, the unclassified event counter is at count four. The next two time periods 670 and 672 are classified as AF based on the corresponding AF scores exceeding the second lower AF score threshold 604. In some examples, if the next time period 670 is also an unclassified time period, such that five unclassified time periods occur consecutively, the AF event counter and the unclassified event counter may be reset to zero, and the AF score threshold may be increased to the first higher AF score threshold 602. The processor 224 may therefore adjust the AF score threshold and reset the counter in response to detecting a predetermined number of consecutive unclassified time periods.

[0137] In the example shown, the next time period 670 is classified as AF, so the unclassified event counter remains at count four. The AF event counter increases to two after time period 670 and increases to three after time period 672. In response to the AF event count reaching the detection threshold (three in this example), AF is detected at time 676. The AF score threshold remains at the second lower threshold 604 until the termination of the AF episode is detected in response to a non-AF classification (e.g., time period 675) or a predetermined number of consecutive unclassified time periods (e.g., five consecutive unclassified time periods). At time 678, the AF score threshold is adjusted to the first higher threshold 602 in response to the non-AF classification of time period 675. As shown by the examples of scenarios (b) and (c), an AF episode can be detected after the AF threshold is adjusted to the second lower threshold.

[0138] The episode duration 655 in scenario (c) begins with the AF period 650 and extends through the AF period 672 resulting in AF detection at time 676. The episode duration 655 includes consecutive unclassified periods 652, 654, 656, and 658 that do not result in the detection of AF termination at 678.

[0139] Scenario (d) shows another example of a series of two-minute time period classifications and corresponding adjustments to the AF score thresholds. An initial time period 680 is classified as AF based on the first higher AF score threshold 602. The AF event counter is incremented to a count of one. A subsequent time period 682 is classified as non-AF based on the first AF score threshold 604. In response to the non-AF classification, the AF event counter may be reset to zero. Since the AF event counter is currently zero, the processor 224 may not count the unclassified time period 684.

[0140] 6. The subsequent AF-U-AF sequence (time periods 686, 688, and 690, respectively) results in an AF event count of two and an unclassified event count of one. Two consecutive time periods 688 and 690, respectively classified as unclassified and AF, following the AF time period 686 result in a combined event count of AF and unclassified time periods equal to three. In response to the combined event count of three, the processor 224 adjusts the AF score threshold from the first higher AF score threshold 602 to the second lower AF score threshold 604 at time 695. Because the AF event count is two, AF has not been detected. Based on a comparison of the AF score with the second lower AF score threshold 604, the next time period 692 is classified as AF. The processor 224 increases the AF event count to three and detects AF at time 697 in response to the AF event count reaching the detection threshold. When AF is detected at time 697, the unclassified event counter is reset to zero. The unclassified event counter will count the unclassified time periods, starting from zero after AF detection, so as to count consecutive unclassified time periods used to detect the termination of the AF episode. The unclassified event count reaches two after time periods 694 and 696. The next time period 698 is classified as non-AF, resulting in the detection of the termination of the AF episode at time 699. At time 699, all event counters are reset to zero and the AF score threshold is adjusted back to the first higher threshold 602.

[0141] The episode duration 685 begins with an AF classification time period 686, which is the earliest AF classification event that results in the detection of AF at time 697. The episode duration includes an unclassified time period 688, which does not contribute to the detected termination at box 699. The episode duration 685 ends with the last AF classification time period 692 before the termination is detected at time 699. In some instances, unclassified time periods 694 and 696 may not be included in the episode duration 685 because they are immediately preceded by a non-AF classified time period 698 that results in the detection of episode termination without an intervening AF classified episode. However, in other instances, those unclassified time periods may also be included in the episode duration 655.

[0142] The first and second AF score thresholds may be fixed values, or may be programmable by a user. In one example, a user may program the first and second AF score thresholds to be increased or set to a relatively higher value than currently programmed, or to be decreased or set to a relatively lower value than currently programmed. In response to a programming command input by the user, both the first and second AF score thresholds are adjusted together upward or downward by the same increment or decrement, respectively. In other examples, a user may programmably select each of the first and second AF score thresholds customized to individual patient needs.

[0143] Although in Fig. 9 Only two different AF score thresholds 602 and 604 are shown in the example of FIG. 69 , but it should be understood that in other examples, the AF score threshold may be adjusted between three or more AF score thresholds. For example, after an AF detection is made at time 697, the AF score threshold set to the second lower threshold 604 may be reduced to a third lowest AF score threshold to allow continued detection of AF episodes using criteria that are less stringent than the initial AF detection criteria. In other examples, once an AF detection is made at time 697, the AF score threshold may be increased from the second lower threshold 604 back to the first higher threshold 602, or to a third intermediate threshold between the second lower threshold 604 and the first higher threshold 602. The third intermediate threshold may be applied to classify subsequent time periods until termination of an AF episode is detected based on a predetermined number of time periods being classified as non-AF (based on the AF score falling below the third intermediate threshold).

[0144] exist Fig. 9 In an example scenario, the AF score threshold is adjusted after at least two consecutive time periods of any combination of AF and unclassified are classified immediately following a preceding or initial AF classification time period. In the example of a two-minute time interval, the AF score threshold is adjusted after six minutes without non-AF classification. In other examples, the AF score threshold is adjusted after fewer or more time periods. For example, a single time period classified as AF may cause the AF score threshold to be reduced to a second lower value. In other words, the AF score threshold may be adjusted immediately in response to time period 610, 630, 650, or 680 being classified as AF (e.g., immediately following a first AF classification time period). In other examples, at least one unclassified or AF time period immediately following a previous AF time period may cause the AF score threshold to be adjusted. In yet other examples, more than two time periods that are not classified as non-AF and that are consecutively followed by a first time period classified as AF may be required before the AF score threshold is adjusted. Thus, for example, the total duration from the first time period classified as AF until the AF score threshold is adjusted may be greater than or equal to 2 minutes and less than or equal to 10 minutes.

[0145] Fig.10 700 is a flow chart of a method for detecting atrial fibrillation according to one example. At block 702, the classification of the current time period is determined. If the classification is AF, as determined at block 704, the processor 224 increments the AF event counter at block 706 and proceeds to block 714 to compare the combined count of the AF event counter and the unclassified event counter to a threshold. If at block 714, the combined count does not meet the threshold, then at block 718 the AF event count is compared to the AF detection threshold.

[0146] If the time period is not classified as AF (the "No" branch of block 704), and is classified as non-AF (the "Yes" branch of block 708), the processor 224 proceeds to block 742. If the AF classification criteria were previously adjusted, they are restored to their initial values. For example, if the AF score threshold was previously adjusted to a second, lower threshold, the AF score threshold is returned to the higher first threshold, as in conjunction with Fig. 9 At block 746, if the AF event counter and the unclassified event counter have been previously incremented to non-zero values, they are reset to zero.

[0147] If the current time period classification obtained at block 702 is neither AF nor non-AF, i.e., if the current time period is determined to be unclassified (the "No" branch of block 708), and the AF counter is not currently active (it has a zero value) (the "No" branch of block 710), the processor 224 determines the classification of the next time period at block 702. If, as determined at block 710, the AF event count is greater than zero, indicating that an initial AF classification has been made, and the current time period is unclassified, the processor 224 increments the unclassified event count by one at block 712. The unclassified event count can be used to control the adjustment of the AF classification criteria before an AF detection is made, such as in conjunction with Fig. 9 If the unclassified event count has reached the predetermined threshold (the "yes" branch of block 713), the processor may restore the initial AF classification criteria (if previously adjusted) at block 742 and reset the unclassified event counter and the AF event counter to zero at block 746. The processor begins again at block 702 with classification for the next time period.

[0148] After increasing the AF event count at box 706 or increasing the unclassified event count at box 712, if the unclassified event count has not reached the predetermined threshold ("No" of box 713), the combined event count can be compared to the threshold at box 714. At box 714, when the combined count of the AF event counter and the unclassified event counter reaches the threshold (e.g., when the combined count is 3), the AF classification criteria can be adjusted at box 716. In one example, the processor 224 adjusts the AF classification criteria by lowering the AF score threshold to a second lower threshold after classifying the first time period as AF based on the first higher AF score threshold as described above and classifying the next two consecutive time periods as any combination of AF or unclassified. Therefore, the AF classification criteria can be adjusted in response to three consecutive time periods being classified as AF, a sequence of AF-U-AF, or a sequence of AF-UU.

[0149] It should be understood that in some examples, once the combined count reaches a predetermined threshold at block 714, and the AF classification criteria have been adjusted before AF detection is made at block 716, the AF classification criteria will not be adjusted again until termination of the AF episode is detected, for example, based on a time period classified as non-AF (block 708) or based on a predetermined number of unclassified time periods (block 713) (e.g., five consecutive unclassified time periods). In other examples, additional adjustments to the AF score may be made before termination of the AF episode is detected, for example, to a third AF score threshold or back to the first highest AF score threshold, as described above.

[0150] At box 718, processor 224 compares the AF event count to the AF detection threshold. When the AF detection threshold has not been reached, processor 224 returns to box 702 to obtain the next time period classification. As described above, after adjusting the AF classification criteria at box 716, if a non-AF classification is made before AF is detected (the "yes" branch of box 708), the AF classification criteria is restored to the initial classification criteria at box 742, and all AF event and unclassified event counters are reset to zero at box 746. If the subsequent time period is classified as AF (the "yes" branch of box 704), the AF event count is increased accordingly at box 706.

[0151] If the AF event count reaches the detection threshold ("yes" branch of block 718), processor 224 detects AF at block 720. An AF detection response is provided at block 721. The response to AF detection may include controlling pacing timing and controlling 212 to deliver atrial antiarrhythmic therapy or to withhold ventricular therapy. The response to AF detection may additionally or alternatively include providing data related to the AF episode (such as time of onset, total duration (as determined from the AV event counter when termination of the AF episode is detected, as discussed below) to the AF event counter, as discussed below, or using Fig. 9 2), storing the onset of cardiac electrical signals in RAM 226 and / or other data related to the AF event in RAM 226. The data may be transmitted to an external device 40 (FIG. 1) for display to a clinician or for communication of the data for use in managing the patient.

[0152] When an AF detection is made at box 720, the unclassified event counter is reset to a count of zero at box 722. The processor 224 may begin counting subsequent time periods determined to be unclassified after an AF detection is made to detect the termination of an AF episode and obtain a next time period classification at box 724. If the next time period classification is AF, as determined at box 726, the AF event count is increased at box 728. The AF classification made after an AF detection is based on the adjusted AF classification criteria. The AF event counter may continue to be increased with each AF classification made after AF is detected at box 720 to determine the duration of the AF episode and determine the AF burden (e.g., the combined duration of all detected AF episodes within a given monitoring interval (such as 24 hours)). Such AF episode data may be transmitted to an external medical device for display or transmitted to a clinician, thereby providing useful information to the clinician when making diagnostic and therapeutic management decisions.

[0153] If the classification of the following time period is not AF (the "No" branch of block 726), but non-AF (the "Yes" branch of block 730), then the termination of the AF episode is detected at block 740. If the classification of the following time period is neither AF nor non-AF (the "No" branch of block 730), that is, if the time period is determined to be unclassified as indicated by block 732, then the unclassified event counter is increased by one at block 734. At block 736, the unclassified event counter is compared to a threshold. If the threshold is not reached, the processor 224 returns to block 724 to obtain the classification of the following time period. If the unclassified count reaches a threshold (e.g., five consecutive unclassified time periods) at block 736, then the termination of the AF episode is detected at block 740.

[0154] If episode termination is detected, the initial AF classification criteria are restored at block 742 and the AF event and unclassified event counters are reset at block 746. The process begins again at block 702.

[0155] Fig.11 800 is a flow chart of a method performed by an ICD or ICD 110 for providing a response to detected AF according to one example. Figure 8-10 ), then at block 804, an AF episode record is stored. The AF event record may include a Lorentz plot or histogram of the RRI data leading to AF classification and AF detection. The episode record stored at block 804 may further include the start time, end time, and total duration of the AF episode. Fig. 9 , shows examples of episode durations 615, 635, 655, and 685 extending from the beginning of the first corresponding AF classification time period leading to AF detection and ending with the last AF classification time period leading to detection of termination of the AF episode. Unclassified episodes leading to termination detection (e.g., unclassified episodes 694 and 696 in scenario (d)) may not be included in the episode duration, while unclassified episodes immediately preceding termination detection (e.g., unclassified episodes 652, 654, 656, and 658 in scenario (c)) are included in the episode duration, e.g., duration 655.

[0156] When an unclassified time period is included in an AF episode or results in the termination detection of an AF episode, the factor(s) that led to the determination of the time period as unclassified can be stored with the AF episode record. For example, if a time period is determined to be unclassified due to detection of other episodes (such as VT or VF), due to oversensing, due to too many short RRIs, or due to too many ventricular pacing pulses during the time period, the factor can be stored to provide the clinician with useful information in diagnosing the patient's cardiac rhythm state for guiding treatment decisions for treating the patient's AF.

[0157] At block 806, processor 224 may determine the patient's AF burden by calculating the total time of AF that was identified over a 24-hour time period (or other predetermined monitoring interval). The calculation of AF burden may include counting or summing all time periods that were classified as AF, or counting or summing only the AF classified time periods that were included in the detected AF episode. The AF burden may also include all unclassified time periods that occurred during the detected AF episode. In some examples, unclassified time periods that were not included in the AF episode duration that led to the termination of the detected AF episode are not included in the AF burden calculation. For example, referring to Fig. 9, scenario (d), unclassified time period 684 is not included in the AF burden calculation because it occurs before the AF episode indicated by episode duration 685. Unclassified time period 688 is included in the AF burden determination because it occurs during the AF episode. Unclassified time periods 694 and 696 are not included in determining AF burden because they result in a termination detection at time 699 and are not included in episode duration 685.

[0158] When AF is detected, the processor 224 may store cardiac signal segments acquired during the detected AF episode. At box 808, the cardiac signal segments are stored in memory with the normal gain of the sense amplifier 200 and the A / D converter 222 used during cardiac signal analysis and processing performed to identify RRI, analyze over-sensed signals, etc. For example, the cardiac signal stored at box 808 can be an EGM signal acquired using the RV coil electrode 28 and ICD housing 15 of Figure 1. Figure 3A In the system 100 of FIG. 1 , the cardiac signal stored at box 808 can be an ECG signal acquired using the defibrillation electrodes 24 or the defibrillation electrodes 126 and the housing 115. A signal stored in a normal range (e.g., having 8-bit resolution sampled at 128 Hz and having an A / D converter input range of +12 mV) can be used to provide an unclipped EGM or ECG signal for morphological analysis (e.g., wavelet template matching) and for storing an unclipped cardiac signal episode in response to detecting a tachyarrhythmia. The normal range signal stored at box 808 can be selected as a far-field or relatively global cardiac electrical signal that is used to generate a display of the electrical rhythm of the patient's heart (which clearly shows the regularity of the R wave morphology and the RRI) for the clinician to see a high-level view of the signal and the corresponding rhythm of the patient. However, depending on the sensing vector, the normal range cardiac signal stored at box 808 may not include an observable or easily observable P wave.

[0159] Thus, when AF is detected, the processor 224 stores a second cardiac signal having a lower range, higher gain setting (e.g., a range of +2 mV, which can be controlled by adjusting the A / D converter input range) in memory at box 810. The high gain, lower range setting provides a clearer view of the P wave in the stored cardiac signal segment when displayed by the external device 40. The high gain, lower range setting may result in clipping of the R wave in the stored second cardiac signal. However, the first cardiac signal stored in the normal range provides a reliable, unclipped display of the R wave morphology.

[0160] A second cardiac signal stored at a high gain, lower range setting may be selected as a second far-field signal or a relatively global signal. Figure 1A or Figure 1B In the system 1, the second cardiac signal can be obtained between the RV coil electrode 24 and the SVC coil electrode 26, or from the AVC coil electrode 26 to the ICD housing 15. Figure 3A In system 100, a second cardiac signal may be acquired using sensing electrode 128 and sensing electrode 131, one of sensing electrodes 128, 130, or 131 paired with one of defibrillation electrodes 124 or 126, or one of electrodes 124, 126, 128, 130, or 131 paired with housing 115.

[0161] When the ICD 10 or ICD 110 receives an interrogation command from the external device 40, the processor 224 controls the telemetry circuit 330 to transmit the AF episode record, the AF burden information, and the stored first normal gain cardiac signal and the second higher gain cardiac signal. The external device is configured to generate a display of the patient's AF data to provide valuable diagnostic information to the clinician to support the treatment decision-making process.

[0162] If AF is not detected (the "NO" branch of box 802), the processor 224 can detect VT or VF at box 820. If VT or VF is not detected (the "NO" branch of box 820), the processor 224 continues to monitor for cardiac tachyarrhythmias at boxes 802 and 820. If AF is not being detected at box 802 but the processor 224 is detecting VT or VF (the "YES" branch of box 820), a VT or VF episode record is stored at box 822, which may include rate, duration, start time, end time, treatment delivered and results, etc. At box 824, a segment of the first cardiac signal with normal gain acquired during the detected episode is stored. The first cardiac signal at normal gain can be a far-field or relatively global signal that provides a high-level view of R wave morphology and RRI as described above, and can be the same signal with the same gain that was stored at box 808 in response to AF detection. At block 826, the third cardiac electrical signal having a normal gain is stored.

[0163] The third cardiac electrical signal may be a near-field or relatively localized signal acquired using a different sensing vector than either of the first or second cardiac electrical signals stored at blocks 808 and 810. For example, the third cardiac signal stored at block 826 may be acquired using a sensing vector different from that of the first or second cardiac electrical signals stored at blocks 808 and 810. Figure 1A and Figure 1B The EGM signal obtained by the RV tip electrode 28 and the RV ring electrode 30 in Figure 3AIn the system 100 of FIG. 8 , the cardiac signal stored at block 826 may be an ECG signal acquired using sensing electrodes 128 and 130. The third cardiac electrical signal is stored at normal gain but is acquired using a near-field or local ventricular sensing vector that may be used to generate a display of a high quality ventricular signal when VT or VF is detected and an atrial tachyarrhythmia is not detected.

[0164] In this manner, the ICD 10, ICD 10', or ICD 110 provides a unique response for storing data depending on whether an atrial tachyarrhythmia is detected, with or without a concurrent SVT, VT, or VF, or whether a ventricular tachyarrhythmia is detected without a concurrent AF detection. The high gain cardiac electrical signals are stored and transmitted to the external device 40 for display to the clinician, providing the clinician with valuable diagnostic information related to the detected AF episode. When an AF episode is detected simultaneously with a ventricular tachyarrhythmia detection, a relationship between the events leading to the two detections can be determined. However, when only a ventricular tachyarrhythmia is detected, the storage and transmission of the near-field or local ventricular signals can provide the clinician with important information about the ventricular rhythm.

[0165] In some examples, processor 224 responds to AF detection by selecting which cardiac electrical signals to store as described above, and transmits a signal having normal gain and range to external device 40. Processor 52 of external device 40 can automatically generate a display on user display 54 that includes a first high gain low range signal for observation of P waves and a second normal gain normal range signal for unclipped observation of R waves.

[0166] Thus, in the foregoing description, apparatus and methods for detecting and responding to atrial tachyarrhythmias have been presented with reference to specific examples. It should be understood that various modifications may be made to the referenced examples without departing from the scope of the following claims, including: modifying the order of steps performed and / or modifying the combination of operations shown in the flowcharts presented herein.

Claims

1. A method for detecting an atrial tachyarrhythmia episode in a medical device, comprising: sensing cardiac signals; Identify the R wave in the cardiac signal that accompanies ventricular depolarization; determining a classification factor from the R-waves identified within a first predetermined time period; classifying the first predetermined time period as an atrial tachyarrhythmia based on comparing the determined classification factor to a plurality of classification criteria, wherein the atrial tachyarrhythmia episode is not detected solely based on the first predetermined time period being classified as an atrial tachyarrhythmia; adjusting a classification criterion in the plurality of classification criteria from a first classification criterion to a second classification criterion by adjusting a threshold value from a first value to a second value less than the first value after classifying the first predetermined time period as an atrial tachyarrhythmia and before detecting an atrial tachyarrhythmia episode; classifying at least one of the subsequent time periods as an atrial tachyarrhythmia by comparing the classification factor determined in the subsequent time period to the adjusted classification criteria; as well as An atrial tachyarrhythmia episode is detected in response to at least one subsequent time period being classified as an atrial tachyarrhythmia based on the adjusted classification criteria.

2. The method according to claim 1, characterized in that Adjusting the classification criteria after the first predetermined time period is classified as atrial tachyarrhythmia comprises: classifying at least the second predetermined time period by comparing the classification factor determined within the second predetermined time period with the classification criterion before adjusting; and The classification criteria is adjusted in response to the second predetermined time period being classified as one of an atrial tachyarrhythmia and unclassified.

3. The method according to claim 1, characterized in that Adjusting the classification criteria after the first predetermined time period is classified as atrial tachyarrhythmia comprises: classifying the plurality of consecutive predetermined time periods before adjusting by comparing the classification factor determined in each of the plurality of consecutive predetermined time periods after the first predetermined time period with a classification criterion; The classification criteria is adjusted in response to none of the plurality of consecutive predetermined time periods being classified as a non-atrial tachyarrhythmia.

4. The method according to claim 1, characterized in that Adjusting the classification criteria includes adjusting the classification criteria in response to the first predetermined time period being classified as an atrial tachyarrhythmia.

5. The method according to claim 1, characterized in that Determining the classification factor includes: determining RR intervals between consecutive R waves in the sensed cardiac signal and determining a variability measure of the RR intervals; Classifying the first predetermined time period includes: classifying the first predetermined time period as an atrial tachyarrhythmia when the variability measure satisfies a threshold; and Adjusting the classification criteria from a first classification criteria to a second classification criteria includes adjusting the variability measure from a first value to a second value that is less than the first value.

6. The method of claim 1, further comprising: The classification criteria is reset from the second classification criteria to the first classification criteria in response to one of at least one subsequent time period being classified as a non-atrial tachyarrhythmia.

7. The method of claim 1, further comprising: The classification criteria is reset from the second classification criteria to the first classification criteria in response to a predetermined number of the subsequent time periods being classified as unclassified.

8. The method of claim 1, further comprising: classifying a second time period after detecting the onset of the atrial tachyarrhythmia; as well as Termination of the atrial tachyarrhythmia episode is detected in response to classifying the second time period as a non-atrial tachyarrhythmia.

9. The method of claim 8, further comprising: The classification criteria is reset from the second classification criteria to the first classification criteria in response to detecting the termination.

10. The method of claim 1, further comprising: classifying a plurality of time periods after detecting the atrial tachyarrhythmia episode using the adjusted classification criteria; detecting termination of the atrial tachyarrhythmia episode in response to a predetermined number of the plurality of time periods being classified as unclassified; and The classification criteria is reset from the second classification criteria to the first classification criteria in response to detecting the termination.

11. The method according to claim 1, characterized in that Determining the classification factor includes: determining whether ventricular tachyarrhythmia detection is occurring during at least one subsequent predetermined time period; and When the ventricular tachyarrhythmia detection is occurring during a subsequent predetermined time period, the subsequent predetermined time period is classified as unclassified.

12. The method of claim 1, further comprising: Responding to detecting the atrial tachyarrhythmia is performed by at least one of: transmitting atrial tachyarrhythmia episode data to another device in response to detecting the atrial tachyarrhythmia, and withholding ventricular therapy.

13. The method according to claim 1, characterized in that Detecting the onset of atrial tachyarrhythmia includes classifying a predetermined number of time periods as atrial tachyarrhythmia, including classifying at least a second time period after the first predetermined time period as atrial tachyarrhythmia based on the first classification criteria, and classifying at least a third time period after the second time period as atrial tachyarrhythmia based on the second classification criteria.

14. The method according to claim 1, wherein: Determining the classification factor includes: determining RR intervals between consecutive R waves in the sensed cardiac signal, and determining a variability measure of the RR intervals having a maximum possible value of 100; classifying the first predetermined time period as an atrial tachyarrhythmia by comparing the variability measure to an atrial fibrillation score threshold; and Adjusting the classification criteria includes adjusting the atrial fibrillation score threshold from the first value to the second value, wherein the first value is at least 25 and equal to or less than 75, and the second value is at least 19 and less than or equal to 60, and wherein the second value is less than the first value.

15. The method according to claim 1, wherein: Classifying at least one subsequent time period includes classifying the at least one subsequent time period as one of an atrial tachyarrhythmia, a non-atrial tachyarrhythmia, or unclassified by comparing a classification factor determined within the subsequent time period to an adjusted classification criterion.

16. A medical device for detecting an atrial tachyarrhythmia episode, comprising: sensing circuitry configured to receive cardiac signals from a plurality of electrodes coupled to the medical device; The processor is configured to: Identify the R wave in the cardiac signal that accompanies ventricular depolarization; determining a classification factor from the R-waves identified within a first predetermined time period; classifying the first predetermined time period as an atrial tachyarrhythmia based on comparing the determined classification factor to a plurality of classification criteria, wherein the atrial tachyarrhythmia episode is not detected solely based on the first predetermined time period being classified as an atrial tachyarrhythmia; adjusting a classification criterion in the plurality of classification criteria from a first classification criterion to a second classification criterion by adjusting a threshold value from a first value to a second value less than the first value after classifying the first predetermined time period as an atrial tachyarrhythmia and before detecting an atrial tachyarrhythmia episode; classifying at least one subsequent time period as an atrial tachyarrhythmia by comparing the classification factor determined during the subsequent time period to the adjusted classification criteria; as well as An atrial tachyarrhythmia episode is detected in response to at least one subsequent time period being classified as an atrial tachyarrhythmia based on the adjusted classification criteria.

17. The device according to claim 16, characterized in that The processor is configured to adjust the classification criteria after the first predetermined time period is classified as an atrial tachyarrhythmia by: classifying at least the second predetermined time period by comparing the classification factor determined within the second predetermined time period with the classification criterion before adjusting; and The classification criteria is adjusted in response to the second predetermined time period being one of not classified as an atrial tachyarrhythmia and not classified.

18. The device according to claim 16, characterized in that The processor is configured to adjust the classification criteria after the first predetermined time period is classified as an atrial tachyarrhythmia by: classifying the plurality of consecutive predetermined time periods before adjusting by comparing a classification factor determined in each of the plurality of consecutive predetermined time periods after the first predetermined time period with a classification criterion; The classification criteria is adjusted in response to none of the plurality of consecutive predetermined time periods being classified as a non-atrial tachyarrhythmia.

19. The device according to claim 18, characterized in that The first predetermined time period and the plurality of consecutive predetermined time periods have a total duration greater than or equal to two minutes and less than or equal to ten minutes.

20. The device according to claim 16, characterized in that The processor is configured to adjust the classification criteria in response to the first predetermined time period being classified as an atrial tachyarrhythmia.

Images