We tested the clinically relevant diagnostic accuracy of a new electrocardiographic (ECG) recording system incorporating all 10 ECG electrodes in a single-size glove worn on the patient’s left arm and placed on the chest. The PhysioGlove (PG) was designed to allow fast, reproducible, electrode placement with only minimal training. The American College of Cardiology/American Heart Association ECG recording guidelines and others have repeatedly highlighted the unacceptable progressive deterioration in ECG accuracy mainly resulting from a performer’s lack of proficiency and diligence, leading to frequent electrode misplacement. We studied 428 consenting adult patients with a broad spectrum of anthropomorphic characteristics and ECG and cardiovascular pathologic entities. The chest girth was measured to ascertain the single-size PG clinical diagnostic accuracy in ≥90% of this patient population. For each patient, a PG and standard-cable electrocardiogram were consecutively recorded and interpreted by experienced electrocardiographers. The study included 3 phases: phase 1, run-in (n = 120); phase 2, comparative diagnostic accuracy (n = 208); and phase 3, randomized, blinded, diagnostic accuracy (n = 100). Of the entire study population (n = 428), 92% fit the chest girth range of 85 to 118 cm (34 to 47 in.), representing the reference standard clinical diagnostic PG chest girth range. The phase 2 PG diagnostic accuracy was 91.3% for entire chest girth range and 95.7% for the 89.4% of patients with a chest girth within the reference range. The mean PG diagnostic accuracy in phase 3 was 93% (95% confidence interval 89% to 95%). In conclusion, compared with standard-cable electrocardiograms, the PG demonstrated excellent diagnostic accuracy (93% to 95.7%) in ≥90% of a typical western adult patient population. The PG’s ease of use and minimal training requirements offer a promising tool to markedly improve ECG clinical diagnostic accuracy in most adult western patients.
The purpose of the present study was to evaluate a new, alternative electrocardiographic (ECG) recording system that simplifies the ECG recording process. This system incorporates all 10 conventional ECG recording electrodes in a glove (PhysioGlove, Commwell, Evanston, Illinois) to be slipped onto the patient’s left arm and placed on the chest ( Figure 1 ). This design has several potential advantages. First, it eliminates the problem of dealing with cumbersome lead wires. Second, no special skin preparation is needed (no hair removal, even plain water provides good electrode skin contact). Third, limb-lead switching is prevented (electrode locations on the glove are fixed). Finally, the adherence to very simple instructions places the precordial leads in their correct and reproducible anatomic positions. All these features could be of utmost importance in improving the accuracy of 12-lead ECG recording in daily practice, even by minimally trained personnel. Moreover, providing the emergency medical service personnel (or even patients or bystanders) with the PhysioGlove (PG) might contribute toward obtaining the earliest possible prehospital 12-lead ECG recording during acute myocardial ischemia or infarction. The aims of the present study were to determine (1) the overall clinical diagnostic accuracy of ECG recordings obtained using the PG compared with those obtained using a standard 12-lead ECG cable in a typical unselected western adult cardiology patient population and (2) the diagnostic accuracy of the single-size PG used in the present study in patients within the reference standard chest girth spectrum, which encompass ≥90% of a typical patient population referred for a diagnostic 12-lead ECG recording.
Methods
We studied a cohort of consenting adult cardiology patients aged 18 to 95 years who had been referred for a diagnostic 12-lead ECG recording. These patients were enrolled at 2 sites: the University of Chicago Medical Center (Chicago, Illinois) and Shaare-Zedek Medical Center (Jerusalem, Israel). Most of the patients were recruited from the cardiology outpatient clinics at the participating hospitals. A small number of inpatients were also recruited. Because of the random nature of the patient enrollment, the present cohort represented a typical population mix in terms of gender, age, race, anthropomorphism, clinical characteristics, and ECG pathologic features. Patients with severe chest malformations, open wounds, or contagious skin eruptions on their chest or left arm were excluded.
The specific size and shape of the study PG was determined using large-scale anthropomorphic population statistics, such as the Civilian American and European Surface Anthropometry Resource (CEASAR) garment industry database, “Size-US,” and “Size-Europe” statistics, the EN-13402-4 garment coding system, and aircrew member anthropomorphic studies. A summary of the United States and European population chest girth distribution data is listed in Table 1 . These extensive population studies have shown that ≥90% of western adults (male and female) will be included in the chest girth range of 82 to 119 cm (32 to 47 in.) corresponding to shirt sizes small to extra-large. The single-size PG was designed to closely fit this population spectrum. A specific algorithm was built into the PG software to compensate for the unconventional location of the limb lead electrodes.
| Name (Convention) | Code | Men (cm) | Women (cm) |
|---|---|---|---|
| Extra-extra-small | XXS | 70–78 | 66–74 |
| Extra-small | XS | 78–86 | 74–82 |
| Small | S | 86–94 | 82–90 |
| Medium | M | 94–102 | 90–98 |
| Large | L | 102–110 | 98–106 |
| Extra-large | XL | 110–118 | 107–119 |
| Extra-extra-large | XXL | 118–129 | 119–131 |
| Extra-extra-extra-large | 3XL | 129–141 | 131–143 |
After the patients provided written informed consent, their demographics, height, weight, and chest girth (at the level of the axilla fold at mid-inspiration) were recorded. The patients’ clinical and ECG diagnoses were retrieved from the corresponding central hospital’s electronic patient data management systems. Each patient had 2 consecutive 12-lead ECG recordings: 1 using the PG, immediately followed by another using a standard ECG cable with the chest electrodes diligently placed in the correct conventional anatomic locations.
The study included 3 phases ( Figure 2 ). In the first two phases, the recordings were made in the order noted, and in the third phase, the order of the recordings was randomized. All ECG tracings were digitally recorded and stored in a personal computed-based ECG system (ES-1 ECG system, Commwell Inc., Evanston, Illinois) by dedicated, specifically trained, research personnel. The PG was thoroughly cleaned between patients using a 70% isopropanol solution according to the Centers for Disease Control and Prevention (Bethesda, Maryland) guidelines. A disposable plastic glove was placed on the patient’s hand before its insertion into the PG.
The run-in phase 1 study consisted of a preliminary training period to acquaint the study team with the protocol and the PG ECG recording system. This phase was also targeted to streamline the performance of phases 2 and 3 of the protocol in the busy environment of the cardiology outpatient clinics at both university hospitals. The phase 1 patient data were included only in the demographic and anthropomorphic analysis of the study population.
In the preliminary, comparative diagnostic accuracy phase (phase 2), the global diagnostic accuracy of the PG-ECG recording system was assessed by a single experienced cardiologist who visually compared side-by-side PG and standard-cable ECG recordings. The reader assessed the clinical diagnostic agreement of the PG recordings, using the standard-cable ECG recording as the reference standard. However, unlike in common routine practice, the expert reader was unaware of the demographic and clinical data of the patients. The final diagnostic data were divided into 2 categories: category 1, clinical diagnostic concordance, when no clinically important differences were noted between the electrocardiograms (examples of minor, clinically nonsignificant, discrepancies included minor ST-T deviation, minor T-wave abnormalities, intraventricular conduction delay); and category 2, clinical diagnostic discordance, with major clinically important differences present between the interpretation of the 2 recordings (examples of major discrepancies included discrepancies in cardiac rhythm, atrial or ventricular hypertrophy, myocardial infarction).
To confirm the basic guiding principles for the design of the single-size PG, we used the anthropomorphic data collected from the phase 1 and 2 patients. The statistical dispersion of the patients’ chest girth allowed the determination of the lower and upper chest girth limits of the reference range, within which about 90% of the study population were included.
In the randomized and blinded phase 3 substudy, an additional 100 patients who fit within the reference clinical diagnostic chest girth range defined in the first 2 phases, were recruited. In the third phase, the 2 ECG recordings (PG and standard cable) were obtained in random order. Three highly experienced cardiologists performed a blinded and separate (not side by side) diagnostic assessment of each ECG recording. For complete blinding, each reader received 4 separate, internally mixed batches of 50 tracings each. Interbatch mixing was performed to prevent inadvertent side-by-side reading of 2 recordings from the same patient. The batches were delivered several days apart to reduce the chances of the reader remembering ECG recordings from one batch to the next. Except for the patient’s code number, the electrocardiograms did not include any other administrative or clinical information, such as the patient’s personal data, demographics, recording mode, sequence of recording, or any other data concerning the electrocardiogram. The ECG diagnostic primary statements and their qualifiers, used by the blinded readers, were based on the latest American Heart Association/American College of Cardiology recommendations. The agreement in the interpretation of the PG and standard-cable ECG recordings for each patient and reader were graded using the same classification as used in phase 2.
In phase 2, diagnostic accuracy was assessed by the proportion of clinically concordant and discordant interpretations (categories 1 and 2) between the 2 recordings. Diagnostic accuracy >95% was regarded as appropriate. The patient choice and sample size in phase 3 was determined from the phase 1 and 2 chest girth statistics. The phase 1 and 2 PG reference standard clinical diagnostic chest girth criteria were statistically adequate to confirm the phase 3, 100-patient sample size, with median diagnostic accuracy of 95% and a confidence interval of ±4.0%. The statistical evaluation of the inter-reader ECG interpretations in phase 3 was performed using the bootstrap approach (resampling patients). The bootstrap data points were a random sample of size n, drawn with replacement, from the sample (x 1 ,…, x n ). Thus, the bootstrap data set consisted of members of the original data set, some appearing 0 times and some appearing ≥1 times. One thousand samples of 100 observations per sample were generated from the original sample, with replacement, using the SAS RANUNI random number generator (SAS Institute, Cary, North Carolina). The middle 95% of the percentage of agreement measure from these 1,000 samples were captured as the 95% confidence interval for the percentage of agreement.
Results
A total of 436 patients were included in the study registry. Of these patients, 8 were excluded because of technical problems (i.e., chest malformations, skin infection). The remaining 428 patients represented the entire study population. The age range was 18 to 95 years, and 155 were women and 273 were men. Of the 428 patients, 184 (43%) were African-American and 244 (57%) were white, Hispanic, or Asian. The study population showed a wide range of demographic and anthropomorphic data. The chest girth distribution for the entire cohort spanned 65 to 155 cm (26 to 61 in.; Figure 3 ). As predicted, 392 of the 428 patients (92%) had a chest girth of 85 to 118 cm (34 to 47 in.). This chest girth spectrum closely fit the predicted (using the population chest girth statistics) PG reference standard target range used to design the single-size study PG.
The spectrum of cardiovascular and other relevant clinical diagnoses for the 308 patients in phases 2 and 3 are summarized in Table 2 . The spectrum of ECG pathologic findings in these patients is summarized in Table 3 . The anthropomorphic data of these patients included a mean height of 171 cm (range 134 to 193 [67 in. range 53 to 76]) and a mean weight of 84 kg (range 41 to 202 [186 lb, range 92 to 445]).
| Diagnosis | n |
|---|---|
| No cardiac pathology | 12 |
| Coronary artery disease | 176 |
| After myocardial infarction | 68 |
| After percutaneous coronary intervention | 148 |
| After coronary artery bypass grafting | 41 |
| After cardiac transplantation | 1 |
| Congestive heart failure | 71 |
| Hypertension | 159 |
| Hyperlipidemia | 89 |
| Atrial fibrillation or flutter paroxysmal, chronic, after ablation, or after maze procedure | 51 |
| Intracardiac defibrillator or pacemaker for atrioventricular block | 18 |
| Valvular disease: aortic, mitral, after aortic or mitral valve replacement | 31 |
| Atrial septal defect or patent foramen ovale closure | 2 |
| Pulmonary hypertension | 5 |
| Substance abuse endocarditis | 6 |
| Chronic renal failure with hemodialysis | 4 |
| Pulmonary embolus | 4 |
| Congenital heart disease | 3 |
| Abdominal aortic aneurysm | 2 |
| Diagnosis | n |
|---|---|
| Normal sinus rhythm | 167 |
| Sinus tachycardia | 31 |
| Sinus bradycardia | 35 |
| Premature atrial contractions | 12 |
| Atrial fibrillation | 45 |
| Atrial flutter | 6 |
| Supraventricular tachyarrhythmia | 3 |
| Ectopic atrial tachycardia, unifocal | 3 |
| Junctional escape complexes | 4 |
| Premature ventricular contractions | 9 |
| Fusion complexes | 6 |
| Accelerated idioventricular rhythm | 3 |
| Prolonged PR interval | 18 |
| Atrioventricular block, advanced (high-grade) | 2 |
| Left anterior hemiblock | 39 |
| Left bundle branch block | 6 |
| Incomplete right bundle branch block | 9 |
| Right bundle branch block | 33 |
| Intraventricular conduction delay | 9 |
| Right axis deviation | 3 |
| Left axis deviation | 15 |
| Left atrial enlargement | 34 |
| Left atrial enlargement (possible or borderline) | 14 |
| Left ventricular hypertrophy | 33 |
| Right ventricular hypertrophy | 3 |
| ST deviation (diffuse probably insignificant) | 6 |
| ST-T deviation (possible ischemia) | 34 |
| ST-T deviation (diffuse, probably insignificant) | 8 |
| T-wave abnormality (probably ischemia) | 49 |
| T-wave abnormality (diffuse, insignificant) | 17 |
| Prolonged QT interval | 9 |
| Prominent U waves | 4 |
| ST-T due to left ventricular hypertrophy | 12 |
| Early repolarization | 15 |
| Abnormal R-wave progression | 33 |
| Anterior myocardial infarction | 6 |
| Inferior myocardial infarction | 33 |
| Anteroseptal myocardial infarction | 18 |
| Extensive myocardial Infarction | 3 |
| All pacing modes | 15 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
