Author + information
- Received August 9, 2000
- Revision received December 8, 2000
- Accepted January 30, 2001
- Published online June 15, 2001.
- Kirk T Spencer, MD, FACCa,* (, )
- Allen S Anderson, MD, FACCa,
- Ajay Bhargava, MDa,
- Amy C Bales, MD, FACCa,
- Matthew Sorrentino, MD, FACCa,
- Kathy Furlong, RNa and
- Roberto M Lang, MD, FACCa
- ↵*Reprint requests and correspondence:
Dr. Kirk Spencer, Noninvasive Cardiac Imaging Laboratory, Department of Cardiology, University of Chicago, 5841 South Maryland Avenue, MC5084, Chicago, Illinois 60637
The purpose of this study was to compare the results of physical examinations (PEs) performed by board-certified cardiologists with the results of point-of-care (POC) echocardiography in a group of patients with cardiovascular disease.
Although cardiovascular PE is crucial in the evaluation of patients with suspected heart disease, the skills required to diagnose abnormal cardiovascular findings have been declining. Echocardiography is a powerful noninvasive cardiovascular diagnostic tool; however, echocardiographic evaluation of patients is not performed at the time of patient encounter (POC echocardiography), beacuse current platforms are cumbersome and expensive for individual physician use. The development of miniaturized echocardiographic equipment has the potential to overcome some of these limitations.
Thirty-six subjects had a complete cardiovascular examination by four board-certified cardiologists. The physicians subsequently imaged each patient using a miniaturized echocardiographic platform. The yield of PE and POC echocardiography were compared using a complete echocardiographic study as the gold standard, performed on an upper-end platform.
Cardiac examination failed to detect 59% of the overall cardiovascular findings. Physician-performed echocardiography with the prototype device missed 29% of the overall cardiovascular pathology. When considering only the major cardiovascular findings, the cardiologists’ PEs still failed to correctly detect 43%. Point-of-care echocardiography reduced this to 21% without significant interphysician variation.
Point-of-care echocardiography using a miniaturized echocardiographic platform substantially improved the detection of important cardiovascular pathology compared with PE. Use of this device by a cardiovascular specialist with training in echocardiography as a routine adjunct to PE appears to be useful.
Cardiovascular physical examination (PE), along with patient interview, forms the cornerstone of the evaluation of the patient with suspected heart disease. A broad spectrum of cardiovascular disorders can be detected and graded with bedside inspection, palpation and auscultation. However, the PE skills required to diagnose abnormal cardiovascular findings have been declining. Studies assessing PE abilities have shown significant error and omission rates for physicians at all levels of training (1). This decline in PE skills has
been attributed to the reliance on newer technological methods of diagnosis as well as decreasing availability of time for bedside teaching (2–4).
Echocardiography is a powerful noninvasive cardiovascular diagnostic tool that provides highly sensitive and specific results. An echocardiography study is typically ordered when the PE or history suggests a cardiovascular disorder. The study is generally performed hours to days after the physician-patient encounter. Echocardiographic evaluation of patients is not performed at the time of this encounter for several reasons. First, there is the belief that PE of the cardiovascular system is sufficiently accurate. In addition, current echocardiographic platforms are cumbersome and expensive for individual physician use. Lastly, substantial training is required to obtain echocardiographic images with a full-feature echocardiographic platform.
The development of miniaturized echocardiographic equipment has the potential to overcome some of the size, expertise and cost limitations of larger platforms (5,6). A small, simplified echocardiographic system, the size of a laptop computer, might allow bedside evaluation of the cardiovascular patient at the time of physician encounter, so called point-of-care (POC) echocardiography. We sought to evaluate whether physician-performed POC echocardiography with a miniaturized platform would provide accurate cardiovascular diagnoses. Specifically, we sought to compare the results of PEs performed by board-certified cardiologists with the results of POC echocardiography in a group of patients with cardiovascular disease.
Thirty-six subjects drawn from the cardiology clinic at the University of Chicago were recruited for participation in this study. Patients were selected based on their cardiovascular diagnoses in order to provide a wide range of clinical findings that could be diagnosed by both PE and echocardiography. In addition, six subjects without known cardiovascular disease were enrolled. Four board-certified cardiologists (with average cardiology practice experience of five years) initially examined and subsequently performed transthoracic echocardiography with the prototype device on each of these subjects (144 examinations). Physicians were provided with a brief clinical history and instructed not to interview the patient. The cardiovascular PE was performed in a quiet office. The exam protocol and use of maneuvers were performed according to each physician’s usual practice, and their findings were recorded immediately afterward.
After completing the PE, physicians imaged each patient using a miniaturized echocardiographic platform (Agilent Technologies, Andover, Massachusetts). This prototype ultrasound device consisted of a base unit (27 × 20 × 8 cm), phased array 2.5 MHz transducer and battery (Fig. 1). The weight of the assembled unit is 6.6 pounds. This platform provides two-dimensional (2D) and color Doppler echocardiographic images on a 14 by 10.5 cm screen. There are a limited number of controls, including those for adjusting imaging depth and gain. Images can be frozen and scrolled for review. An electronic caliper and touch pad allow measurements to be performed.
No preset ultrasound exam protocol was followed, and imaging was limited to ≤15 min in duration. All physicians had been trained in the use of this device and had performed 10 to 15 examinations before study initiation. The physicians all had level II training in the performance and interpretation of echocardiographic images. Level II training, as defined by the American Society of Echocardiography, requires performing and interpreting a minimum of 300 echocardiographic examinations. After completion of the echocardiographic study, the physicians again recorded their cardiovascular findings. Examples of 2D echocardiographic still-frames acquired with the prototype ultrasound device in patients enrolled in this study are depicted in Figures 2 and 3. ⇓⇓
All subjects had a complete echocardiographic study with an upper-end platform (Sonos 5500, Agilent Technologies, Andover, Massachusetts), which served as the gold standard for comparison. A sonographer with at least five years of cardiovascular experience who did not have knowledge of the miniaturized echocardiographic results performed these exams. Standard views were obtained, and digital loops were recorded. No special maneuvers or agents were used to enhance the ability to detect abnormalities. A cardiologist not involved in the patient evaluation interpreted these studies. The presence of mild or more severe than mild valvular regurgitation, valvular stenosis, or right ventricular (RV) or left ventricular (LV) systolic dysfunction was noted. In addition, the presence of mitral valve prolapse or ventricular septal defect was recorded.
Cardiovascular findings noted on the complete echocardiographic study were recorded and divided into major and minor diagnoses depending on their clinical importance. Major findings were considered those that would result in further diagnostic evaluation, alter the patient’s prognosis, require pharmacological therapy or indicate endocarditis prophylaxis. Specifically, this included moderate or more severe than moderate valvular regurgitation or stenosis, moderate or worse ventricular dysfunction, hypertrophic cardiomyopathy, mitral valve prolapse or ventricular septal defect.
The percentage of findings present on the gold standard echocardiographic examination that were also noted by PE or POC echocardiography was computed for each physician as well as for all observers combined. Differences between physicians were compared with a chi-squared test. The yields of PE and POC echocardiography were compared with a McNemar chi-squared test. The detection rates of individual findings for echocardiography and PE were compared with a chi-squared test with Yates’ correction.
There was a total of 79 cardiovascular findings detected by the gold standard echocardiographic examination in this patient population. Of these 79 findings, 34 were considered to be of major clinical importance. Imaging with the device was easily learned and readily performed in the office setting. Representative images are shown in Figures 2 and 3.
Cardiac examination failed to detect 59% of the overall cardiovascular findings, with all four physicians having similar rates of exam failure (Table 1). Physician-performed echocardiography with the prototype device missed 29% of the overall cardiovascular pathology, without significant difference among the four physicians. This 51% reduction in unnoted cardiovascular abnormalities was statistically significant. When considering only the major cardiovascular findings, the cardiologists’ PEs still failed to correctly detect 43% (Table 1). Point-of-care echocardiography reduced this to 21% without significant interphysician variation. This represents a statistically significant 50% increase in the detection of major abnormalities during the patient encounter.
As shown in Table 2, only 3% of the findings, whether major or minor, were detected by PE and missed by POC echocardiography. Both techniques considered together missed 26% of overall and 18% of major diagnoses. Among the normal patients, three of six of the PEs revealed an abnormality not present on the gold standard echocardiographic study. There was a total of six false positive PE findings including: mitral regurgitation (2), mitral valve prolapse (2), tricuspid regurgitation (1)and LV dysfunction (1). Point-of-care echocardiography was abnormal in two of six normal subjects, with four findings: two mild mitral regurgitation, one mild tricuspid regurgitation and one mitral valve prolapse.
Among the valvular findings, there was a significant increase in the detection of systolic murmurs with POC echocardiography over clinical examination (Table 3). Aortic stenosis, mitral regurgitation and tricuspid regurgitation were correctly detected two to four times as often with physician-performed echocardiography compared with PE. There was also an increase in diastolic valvular lesion detection with POC echocardiography that was primarily due to improved aortic insufficiency diagnosis.
Physical examination identified major valvular lesions with a similar accuracy to that of POC echocardiography with the exception of mitral stenosis, which was never missed by echocardiography (Table 4). Echocardiography was superior to PE for the identification of significant ventricular systolic dysfunction and mitral valve prolapse (Table 4). Among the two patients with ventricular septal defect, each patient was misidentified by at least one of the cardiologists whereas, using echocardiography, all four physicians identified both patients’ ventricular septal defect.
PE for the evaluation of cardiovascular disease
Physical examination is one of the essential components of patient evaluation in medicine. In addition to the diagnostic value of patient examination, it has served as an important discriminator for the subsequent referral for more sophisticated, and, thus, more costly, evaluation. However, the PE skills of physicians are declining. Among the component parts of a complete patient examination, cardiovascular exam skills have been shown to be particularly poor (7). Although cardiac auscultation has excellent specificity, the sensitivity of this technique in the current era is remarkably low (1,4,8–13).
Cardiologists in our study identified mitral and tricuspid regurgitation murmurs in 43% and 24% of subjects, respectively. These results are comparable with previously published studies in which these murmurs were detected in 13% to 56% and 28% to 33% of patients, respectively (1,4,9–11). Our physicians correctly identified the murmur of aortic stenosis more often (88%) compared with prior studies in which this lesion was identified in only 20% to 54% of patients (4,10,14). The PE for the evaluation of aortic regurgitation and mitral stenosis is even poorer, with published sensitivities for the detection of diastolic murmurs as low as 5% to 24%, which are similar to our result of 26% (4,8,11,13).
Despite the clinical importance of identifying patients with ventricular dysfunction, our study and others have demonstrated a limited ability for PE to diagnose decreased RV and LV systolic performance (15). In this study, nearly two thirds of patients with ventricular dysfunction went undetected with clinical evaluation alone. The poor sensitivity for PE findings for the detection of LV dysfunction has been previously noted (15). Although several exam findings are suggestive of ventricular dysfunction, they are often difficult to illicit, and their overall sensitivity and specificity do not allow reliable assessment of ventricular function (15).
Echocardiography for the evaluation of cardiovascular disease
Echocardiography often serves as the gold standard in studies evaluating the accuracy of cardiac auscultation. It has been speculated that cardiac ultrasound is, in fact, “too sensitive” in that it detects lesions that are of negligible clinical importance. In this study, we omitted patients with trivial valvular abnormalities. In addition, we stratified findings according to clinical importance and showed that PE was inaccurate even for abnormalities that would alter diagnostic evaluation, therapeutic plan or patient prognosis.
Prior studies have demonstrated that the yield of PE improves with the increasing severity of valvular disease. However, in this study, even when considering only the more severe grades of aortic stenosis, mitral regurgitation and tricuspid regurgitation, the PE failed to correctly identify these lesions in 28%, 23% and 35% of patients, respectively. These data are consistent with previous studies demonstrating high rates of missed diagnosis for higher grades of systolic murmurs (14% to 50% [8,9,16]).
Despite the clear superiority of echocardiography over PE, it is not used routinely during physician-patient encounters. One of the major impediments is the high cost of high-end echocardiographic systems, which typically exceed $200,000. In addition, current echocardiographic platforms are too cumbersome for practical bedside use. Moreover, in most U.S. hospitals, sonographers, rather than physicians, perform the echocardiographic examination, which makes it impractical to perform POC echocardiography.
In most laboratories, outpatient echocardiographic studies are performed several days after the patient-physician encounter. This results in a percentage of patients failing to return for the test and requires an inconvenient additional trip to the facility for the returning patient. The report of an echocardiographic test also takes days to return to the referring physician. These delays preclude completion of the diagnostic evaluation and initiation of therapy at the time of the patient visit.
This study supports the diagnostic utility of POC echocardiography, that is, cardiac ultrasonography at the time of the PE performed by physicians with prior training in echocardiography. The miniaturized ultrasound unit, when used by physicians, proved to be superior to PE across all cardiovascular pathologies, including stenotic lesions, regurgitant lesions and ventricular dysfunction. The prototype device’s size and weight are clearly practical for carrying while making rounds or evaluating patients in a clinic. Because these platforms will prove to be substantially cheaper than current ultrasound devices, they could be owned by individual physicians, allowing immediate and accurate evaluation of their patients’ cardiovascular system. As with other technologies, reimbursement may be a factor in the future use of these devices.
Although ultrasound is clearly superior for the detection of many cardiovascular lesions, it should not replace the cardiac examination. There are several components of a cardiovascular examination that yield diagnostic information that is not directly detected with echocardiography, such as jugular pulse, peripheral arterial examination, pedal edema and ventricular gallops. In addition, not all patients have adequate acoustic windows to allow diagnostic echocardiographic exams. If the device and exam were both routinely performed, they would likely be complementary. Immediate feedback from the echocardiographic device would, undoubtedly, only serve to improve cardiac examination skills.
Use of this device by physicians requires training in the performance and interpretation of echocardiographic studies. The physicians who participated in this study fulfilled three years of cardiovascular education, which included four to six months of echocardiography. In a typical cardiology fellowship program, this would include the performance and interpretation of several hundred echocardiograms. Use of this equipment by physicians without adequate training could result in important omissions and misdiagnoses, with potential adverse clinical implications. Whether cardiologists with more years of clinical experience or trained at a time when greater emphasis was placed on bedside examination skills would find this device to be helpful is unknown. However, our four physicians were university trained, board-certified cardiologists and likely represent the clinical acumen of the new generation of American cardiologists.
The portable device, when used by cardiologists, still missed important cardiovascular findings. The reason for this is likely multifactorial. Adult cardiologists in the U.S. do not routinely perform echocardiograms; and although they are experienced in the interpretation of studies, they have less experience with data acquisition. In addition, more experience with this specific device might improve diagnostic yield. Another limitation is that the color Doppler capability of the prototype device used in this study was initially not optimal and has been improved. Lastly, these devices simply do not provide image quality identical to that of a high-end platform.
The prototype echocardiographic platform has limited capabilities compared with upper-end platforms, which may also have played a role in its failure to identify certain cardiovascular lesions. Because of the limited imaging capabilities, POC echocardiography cannot replace complete echocardiographic studies. The role of this device in clinical practice should be that of a screening instrument similar to the stethoscope. Patients with abnormal findings should be referred for complete echocardiographic evaluation, which would provide better image quality and complete hemodynamic evaluation.
☆ We thank Susan Moynihan and Toni Burkett from Agilent Technologies (Andover, Massachusetts) for their support. We also thank Agilent Technologies for providing the equipment used in this study.
- left ventricle
- physical examination
- right ventricle
- Received August 9, 2000.
- Revision received December 8, 2000.
- Accepted January 30, 2001.
- American College of Cardiology
- Pritchett A.M,
- Bruce C.J,
- Bailey K.R,
- Tajik A.J,
- Seward J.B
- Pandian N.G,
- Ramasamy S,
- Martin P,
- Banerjee A
- Jaffe W.M,
- Roche A.H.G,
- Coverdale H.A,
- McAlister H.F,
- Ormiston J.A,
- Greene E.R