Luc Dekie, PhD is the Director of Scientific Affairs, Europe.

Luc Dekie

	Atrial fibrillation (Afib) is the most common arrhythmia. According to the Centers for Disease Control and Prevention website, between 2.7 and 6.1 million people in the US have Afib and the incidence increases with age.

 The arrhythmia is associated with a five-fold risk of stroke and a three-fold incidence of congestive heart failure, and higher mortality.

 It is caused by multiple chaotic re-entry circuits in the atria which results in an irregular atrial rhythm and an irregular ventricular rhythm. Atrial fibrillation can occur without a clear cause or can be secondary to cardiovascular pathology or to non-cardiac causes. Atrial fibrillation is classified as follows:

			Paroxysmal: Afib occurs in episodes and terminates spontaneously within seven days

			Persistent: continuous episodes of Afib that last more than seven days

			Longstanding persistent: continuous Afib with a duration of longer than 12 months

			Permanent: the patient and clinician make a joint decision to stop further attempts to restore and/or maintain sinus rhythm. Acceptance of Afib represents a therapeutic attitude on the part of the patient and clinician rather than an inherent pathophysiological attribute of Afib. Acceptance of Afib may change as symptoms, efficacy of therapeutic interventions, and patient and clinician preferences evolve.

		Most patients with paroxysmal Afib eventually develop permanent Afib.

		On an electrocardiogram (ECG), Afib is characterized by an irregular rhythm with a rate around 100-160 beats per minute (bpm) but slower if on medication. The atrial rate is typically 300-500 bpm. The QRS duration is usually normal unless there are additional conduction abnormalities and the P waves are not distinguishable as the atria are firing from multiple locations. They are replaced by multiform fibrillatory waves causing the baseline to be undulating (Figure 1).

		When recording ECGs in the clinic it is of the utmost importance to collect tracings of the best possible quality since signals with noise and artefact might have the appearance of atrial fibrillation and will be diagnosed as such by the ECG machines. Various publications have shown that not all primary physicians are able to pick up the false interpretation. In a study performed by F. Bogun et al., 24% of the patients who were wrongly diagnosed with Afib by the device algorithm were treated for Afib by the physician who was not able to correct the inaccurate interpretation. This resulted in a change in patient management and initiation of inappropriate treatment, including antiarrhythmic medications and anticoagulation as well as unnecessary additional diagnostic testing.

 Other studies have shown that automated Afib diagnosis does get overruled correctly by physicians.

		When recording a 12 lead ECG it is very important to look at all twelve leads before making a final interpretation. The ECGs shown in figure 2 were labeled as atrial fibrillation by the machine algorithm. Careful review of the tracings, however, shows the presence of P wave prior to the QRS complexes in some of the leads and rules out Afib. Several sources indicate that the inaccuracy of ECG algorithms to detect Afib is around 20%.

			When Afib follow up is performed using cardiac event monitors that typically record only one channel it can be even more difficult to diagnose Afib. Figure 3 shows a 16s recording of a 74 year old male patient. The baseline is showing significant artefact mimicking multiform fibrillatory waves. The first 8s show an irregular rhythm, which could be considered to be Afib. Careful review shows regular P waves with constant PQ duration before some of the QRS. These are more visible in the second strip. This patient’s rhythm was therefore considered to be sinus rhythm with conducted isolated atrial premature contractions. In a study performed by M. H. Bae et al., who looked at the underlying rhythm in patients with over-interpretation of Afib, sinus rhythm with atrial premature complexes was the most prominent missed interpretation by the computer algorithm followed by sinus rhythm with artefact.

				To conclude, correct detection of Afib is dependent on good quality recordings. These are important to guide the physician in prescribing the correct treatment to the patients. ECG machines provide an automatic interpretation that should be used as an indication but should not be considered to be the gold standard. ECGs should be over read by physicians who have received the correct training to detect abnormalities like Afib or should be sent to an expert in a hospital or central reading facility.

, PhD, is the Director of Scientific Affairs, Europe at Biomedical Systems.

http://www.cdc.gov/dhdsp/data_statistics/fact_sheets/fs_atrial_fibrillation.htm

A. J. Camm et al. 2012 focused update of the ESC Guidelines for the management of atrial fibrillation. European Heart Journal (2012) 33, 2719–2747.

C. T. January et al. 2014 AHA/ACC/HRS Guideline for the Management of Patients With Atrial Fibrillation: Executive Summary: A Report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines and the Heart Rhythm Society. J Am Coll Cardiol. 2014;64(21):2246-2280. 

F. Bogun et al. Misdiagnosis of Atrial Fibrillation and Its Clinical Consequences. Am J Med. 2004; 117:636–642

Davidenko JM, Snyder LS. Causes of errors in the electrocardiographic diagnosis of atrial fibrillation by physicians. J Electrocardiol 2007;40:450–6.

M. H. Bae et al. Erroneous Computer Electrocardiogram Interpretation of Atrial Fibrillation and Its Clinical Consequences. Clin. Cardiol. 35, 6, 348–353 (2012)

Articles

This article discusses atrial fibrillation (Afib) detection and ECG recordings of the arrhythmia.

Afib Detection: True or False Due to Bad Quality Recordings

	Electrocardiograms (ECGs) are frequently recorded in the clinic as an easy-to-perform, non-invasive method to collect diagnostic information about the status of the heart (rhythm, conduction, chamber enlargement, ischemia, infarction, the effect of electrolyte abnormalities and drug effects, and even non-cardiac disease effects on the heart). Properly recorded ECGs are very reproducible, and can be used serially to determine change in cardiac status. Recording a good quality ECG is straightforward as long as the proper technique is used. This test is usually performed by technicians, physicians, and nurses who have received the proper training to collect high quality recordings.

	In clinical trials, ECGs are usually part of the safety assessments regardless of the therapeutic area. As a result, ECGs are sometimes recorded by nurses and technicians who lack recent training and who are unaware of all the factors that are important to record ECGs of high standard. Quantitative data from the recordings, and even computer generated interpretive statements, may be automatically transferred to trial databases or be used in case report forms. Unless an ECG is unreadable or technical statements are routinely included during the ECG review, little is known about the quality of the ECG when the database is examined.

	This article will discuss the importance of good patient preparation, lead hook-up, and the quality checks that can be performed to ascertain that the ECG is of good quality.

	High quality patient preparation is the key to collecting high quality ECGs. A simple patient preparation procedure tends to work best and is easily taught. The following steps are required for a good patient preparation.

		If an ECG is to be recorded at a visit, the patient can be informed ahead of time not to use lotion or cream on the chest or limbs the day of the recording. Lotions and creams can affect skin prep and electrode adhesion.
		 

		Allow adequate time for the ECG procedure to be performed. Ten to fifteen minutes is required to prepare the patient and record the trace. Do not schedule other procedures, such as blood draws, concurrent with the ECG.
		 

		Ensure the patient is comfortable and tell them to completely relax arms and legs at the time of acquisition. If a patient is not relaxed, the ECG recording will contain the cardiac activity as well as high frequency muscle potentials and low voltage interference. Some patients might not be able to relax due to their disease. For COPD patients, a semi-sitting position may be better than lying flat, and for Parkinson's disease patients it has been proposed to have them put their hands under their body.
		 

		Some patients may be anxious if they have not had an ECG performed before. Explaining to them the procedure will make them feel confident and reassured that there will be no pain or discomfort.
		 

		The room should not be too hot or cold, and the bed should be wide enough to allow the patient to rest and have their arms lie comfortably next to their body.
		 

		Access to the patient’s torso, arms, and lower legs is crucial for correct electrode application. This implies that the patient will have their chest exposed during lead placement. Privacy should be maintained when possible and can be done by covering the patient with a blanket once the leads are connected.
		 

		Ensure a good skin preparation to assure there is a good connection between the electrodes and the skin. Lightly abrading the skin using a sterile gauze is usually sufficient to clean the skin and also helps to reduce skin impedance. Before applying the electrodes shave hairy skin, rub electrode sites with an alcohol swab if needed, and abrade the skin, then allow the skin to air-dry before applying the electrodes. Alcohol dries out the skin, which increases impedance-this should be avoided, if possible.
		 

		Check electrode expiration date and make sure the electrodes are moist. Be aware that electrode packages that are left open to the air can dry the electrodes and result in poor contact. Connect the electrodes to the lead wires and then put the electrodes on the patients. Make sure the electrodes are connected well to the skin while avoiding pushing on the central area of the electrode that contains the conducting gel, considering this can cause the gel to leak, trap air, reduce the adhesion and conduction of the electrical signal, which all can increase noise.
		 

		Make sure that the wires are not pulling on the electrodes and that there is no traction or tension on the wires.
		 

		There should not be any distractions or unnecessary activity in the room, including TV or video games.
		 

		Ensure electrodes are correctly placed.

	ECG electrodes are typically placed using the Mason-Likar lead placement or the standard lead placement scheme. Although the rhythm analysis is not adversely affected by the Mason-Likar configuration, many studies have shown significant differences in morphology between ECGs acquired with the limb electrodes in the standard position and in the Mason-Likar position, and clinicians should be warned that ECGs acquired with the electrodes in the Mason-Likar configuration are not equivalent to those taken with the electrodes on the legs and arms and should not be directly compared.

	2.1.1          Precordial lead placement

	Six electrodes are placed on the chest in the following locations:

			V1, fourth intercostal space at the right sternal border

			V2, fourth intercostal space at the left sternal border

			V4, fifth intercostal space in the midclavicular line

			V5, in the horizontal plane of V4 at the anterior axillary line, or if the anterior axillary line is ambiguous, midway between V4 and V6

			V6, in the horizontal plane of V4 at the midaxillary line

		V1 is the important starting point since the V2-V6 electrode placement will be placed relative to the V1 electrode. It can be difficult locating the fourth intercostal space. The best way is to run your fingers down the sternum, starting at the heads of the clavicles, until you meet a bony horizontal ridge (the sternal angle or angle of Louis). This is easier to find in male patients. With your finger on this ridge, slide it to the patient's right, your finger will drop into an intercostal space, this is the second intercostal space; now move down to the third and then the fourth, that's where you place V1.

		A common error is superior misplacement of V1 and V2 in the second or third intercostal space. This can result in a reduction of initial R-wave amplitude in these leads, approximating 0.1 mV per interspace, which can cause poor R-wave progression or erroneous signs of anterior infarction. Superior displacement of the V1 and V2 electrodes will often result in rSr' complexes with T-wave inversion, resembling the complex in lead aVR.

 Misplacement of V1 and V2 electrodes too high or too low also results in misplacement of the other chest leads, often causing diagnostic errors in the interpretation of left ventricular hypertrophy.

			2.1.2          Limb lead placement

			Arm lead electrodes are placed on the wrists. Leg lead electrodes are placed approximately 7-12 cm (3-5 inches) above the ankles.

			2.2       Mason-Likar lead placement

			2.2.1          Precordial lead placement:

			2.2.2          Limb lead placement:

			Arm electrodes are placed 2 cm below the lateral end of each clavicle, and the left leg electrode is placed in the anterior axillary line halfway between the anterior iliac spine and the costal margin. The ground electrode can be placed anywhere on the body.

				Filtering of the ECG signal is used to reduce noise in the raw signal. The ECG signal is usually low (around 1 mV) so it is important that the filtering does not distort the ECG waveform. The noise may come from muscle noise, electrical interference (50/60Hz), electrical noise from equipment in the environment, and from within the ECG equipment itself, such as from internal AC/DC converters.

				The low frequency noise typically corresponds to the baseline wandering produced by respiration, but baseline drift suppression by filters may affect the low frequency components of the ECG signal, namely the ST-T segments, and may produce artefactual ST-segment deviation.

				Reduction of high frequency noise (typically muscle noise) by digital filters may affect the high frequency component of the ECG signal (e.g. QRS complex). Inadequate high frequency response reduces the amplitude of QRS measurements and the ability to detect small deflections.

				In general, low-pass filters are used to reduce high frequency artefact, while high-pass filters are used to correct for low-frequency artefact. The notch filter is used to reduce electrical artefact of a specific frequency (e.g. 50 or 60 Hz interference from power sources).7,8

				The international guidelines suggest to record the ECGs with a low-frequency cut-off at 0.05 Hz for routine filters and an upper frequency cut-off of at least 150 Hz (250 Hz being more appropriate for infants).

 An obvious consequence of these high frequency recommendations is that reduction of noise by setting the high frequency cut-off of a standard or monitoring ECG to 35 or 40 Hz will invalidate any amplitude measurements used for diagnostic classification 

					It is highly recommended to perform ECG recordings without additional filters, and not change filter settings between recordings on the same patient.

					Most ECG machines have a screen that shows the ECG signal prior to acquiring and printing the actual ECG. This feature should be used by the site to check the tracing for poor quality before pressing the record button. At all times, it is crucial to carefully review the recorded ECG on paper or on screen for completeness and good quality, before unhooking the patient, so that another ECG can be taken if needed.

						Assure correct lead placement: A very common mistake is reversal of the left and the right arm leads. This occurs in about 3% of the ECGs recorded in a hospital setting.

 Consider this especially when the ECG tracing appears questionable, with markedly negative QRS and inverted P and T-waves in lead I and positive P-waves and QRS complexes in Lead aVR. Always check the lead placement if the P-wave in lead I is negative. A flat line (pseudo asystole) seen in lead II or III can be due to a misplacement of the right leg lead with an arm lead.
						 

						Eradicate or minimize any muscular noise: This is the most frequent quality issue. It presents usually like mild or moderate, markedly irregular baseline wavelets, notching, or spikes 

. If the patient is shivering or clinching his/her fists or putting tension on other muscles, it will cause high frequency noise. This kind of noise can be considered as cardiac beats by the ECG device algorithm, which will result in incorrect cardiac interval measurements and misinterpretation. If present, check good adherence of all electrodes, especially the neutral lead. Double check that the patient is lying relaxed and comfortable, and ask the person to lie down as quiet as possible and not to speak.
						

						Check the baseline: it must be as straight and horizontal as possible, without up- and down-sloping or drift 

. If it is not, try to have the patient comfortably lying or sitting, as quiet as possible, breathing smoothly. If possible, ask the patient to hold his breath during the ECG recording. Avoid any lead wire movement and oscillation.

							Look at possible AC electrical interference: It appears like a high frequency, regular sinusoidal or saw tooth pattern, or spiky oscillation 

. Laptops connected to the power supply can also generate an electromagnetic field. A laptop close to the acquisition module can, therefore, also generate high frequency noise artefact. Some mobile phones might also create electrical interference. In case of recurrent problems at the same site, consider a possible defect of the building's electrical installation earth wiring. If possible, record an ECG using the internal battery. Switch off all non-essential equipment in the room, and if fluorescent lights are in use, try turning them off. Consider recording an ECG in a different room, when in doubt, if quality issues are due to room setting.

This article discusses the importance of good patient preparation, lead hook-up, and the quality checks that can be performed to ascertain that an ECG is of good quality.

Luc Dekie

Articles

Afib Detection: True or False Due to Bad Quality Recordings

High Quality ECG Recording Assurance