Does This Patient Have Aortic Regurgitation?

THE RATIONAL CLINICAL EXAMINATION

Niteesh K. Choudhry, MD Objective To review evidence as to the precision and accuracy of clinical examina- Edward E. Etchells, MD, MSc tion for aortic regurgitation (AR). Methods We conducted a structured MEDLINE search of English-language articles CLINICAL SCENARIO (January 1966-July 1997), manually reviewed all reference lists of potentially relevant You are asked to see a 59-year-old articles, and contacted authors of relevant studies for additional information. Each study woman with liver cirrhosis who will be (n = 16) was independently reviewed by both authors and graded for methodological undergoing sclerotherapy for esopha- quality. geal varices. When she was examined by Results Most studies assessed cardiologists as examiners. Cardiologists’ precision for her primary care physician, she had a detecting diastolic murmurs was moderate using audiotapes (␬ = 0.51) and was good pulse pressure of 70 mm Hg. The pri- in the clinical setting (simple agreement, 94%). The most useful finding for ruling in mary care physician is concerned about AR is the presence of an early diastolic murmur (positive likelihood ratio [LR], 8.8- the possibility of aortic regurgitation 32.0 [95% confidence interval {CI}, 2.8-32 to 16-63] for detecting mild or greater AR (AR) and asks you whether endocardi- and 4.0-8.3 [95% CI, 2.5-6.9 to 6.2-11] for detecting moderate or greater AR) (2 tis prophylaxis is necessary for sclero- grade A studies). The most useful finding for ruling out AR is the absence of early diastolic murmur (negative LR, 0.2-0.3 [95% CI, 0.1-0.3 to 0.2-0.4) for mild or greater therapy. You conduct a complete physi- AR and 0.1 [95% CI, 0.0-0.3] for moderate or greater AR) (2 grade A studies). Except cal examination and hear no early for a test evaluating the response to transient arterial occlusion (positive LR, 8.4 [95% diastolic murmur in the third or fourth CI, 1.3-81.0]; negative LR, 0.3 [95% CI, 0.1-0.8]), most signs display poor sensitivity intercostal spaces at the left-sternal bor- and specificity for AR. der. You feel that the patient is unlikely Conclusion Clinical examination by cardiologists is accurate for detecting AR, but to have AR and that endocarditis pro- not enough is known about the examinations of less-expert clinicians. phylaxis is not needed. You suspect that JAMA. 1999;281:2231-2238 www.jama.com the wide pulse pressure is a peripheral hemodynamic consequence of cirrho- creased in recent years. It is estimated ing blood viscosity, blood flow veloc- sis, not AR. The primary care physi- that 2% of the general population un- ity and turbulence, the distance between cian, however, wonders whether the pro- dergoes noninvasive cardiac diagnos- the vibrations and the stethoscope, the cedure should be delayed until an tic evaluation annually.3 If a careful angle at which the vibrations meet the echocardiogram can be obtained. clinical examination can exclude the stethoscope, the transmission quali- presence of AR, then there would be no ties of the tissue between the vibra- WHY IS THE CLINICAL need to proceed with further cardiac tion and the stethoscope, and the au- EXAMINATION IMPORTANT evaluation. ditory capabilities of the examiner.5 IN EVALUATING FOR AR? Aortic regurgitation is a potentially se- Anatomical and Physiological How to Examine for AR rious cardiac abnormality that may be Origins of Diastolic Murmurs A complete clinical history and physi- caused by important underlying disor- The cardinal manifestation of AR is a cal examination are essential in the ders. Patients with AR require careful diastolic murmur. Diastolic murmurs Author Affiliations: Division of General Internal Medi- clinical monitoring to identify the op- are important indicators of structural cine and Clinical Epidemiology, Department of Medi- timal time for surgical intervention. cardiac abnormalities or pathological cine, University of Toronto and the University Health Asymptomatic patients with severe AR states of increased flow (TABLE 1). As Network, Toronto, Ontario. 1 Corresponding Author: Edward E. Etchells, MD, MSc, may benefit from vasodilator therapy. discussed in a previous article in this University Health Network, 200 Elizabeth St, eng- Endocarditis prophylaxis may be indi- series,4 heart murmurs are produced 248, Toronto, Ontario, Canada M5G 2C4 (e-mail: [email protected]). cated for patients with AR who are un- when turbulent blood flow causes pro- The Rational Clinical Examination Section Editors: dergoing various procedures.2 longed auditory vibrations of cardiac David L. Simel, MD, MHS, Durham Veterans Affairs Medical Center and Duke University Medical Center, The use of noninvasive cardiac test- structures. The intensity of the mur- Durham, NC; Drummond Rennie, MD, Deputy Edi- ing, such as echocardiography, has in- mur depends on many factors, includ- tor (West), JAMA.

Downloaded from www.jama.com at University of Florida, on May 24, 2005 AORTIC REGURGITATION evaluation of patients with a diastolic The precision and accuracy of many systolic murmurs but may also be used murmur. A diastolic murmur in a pa- individual components of the examina- to describe diastolic murmurs. A grade tient with renal failure and volume over- tion for AR, including all of the cardiac 1 murmur is not heard immediately on load will have different significance than history and most of the physical exami- auscultation, but is heard after the ex- a diastolic murmur in a patient with a nation, have not been adequately evalu- aminer focuses for a few seconds. Grade history of rheumatic fever and atrial fi- ated. This article will focus on aspects 2 murmurs are heard immediately on brillation. of the cardiac physical examination that auscultation but are softer than the loud The examiner’s ability to detect a di- have been sufficiently assessed for pre- grade 3. Grade 4 murmurs are associ- astolic murmur can be undermined by cision or accuracy. ated with a palpable precordial vibra- environmental factors such as noisy tion called a thrill. (Grade 5 and 6 mur- rooms, examiner factors such as fa- Cardiac Auscultation murs are also associated with a thrill. A tigue or haste, and patient factors such During routine auscultation, the ex- grade 5 murmur is audible when only as dyspnea or tachycardia.6 If examin- aminer attempts to detect a diastolic 1 edge of the stethoscope is on the chest, ing conditions are not optimal, the ex- murmur. Diastole is the period that be- and a grade 6 murmur is audible with amination should be repeated when gins with the closure of the aortic and the entire stethoscope lifted off the conditions improve. pulmonic valves (second heart sound chest.) [S2]) and ends with the closure of the The typical murmur of AR is an early mitral and tricuspid valves (first heart Table 1. Selected Causes of Diastolic diastolic, decrescendo blowing sound Murmurs sound [S1]). A common maneuver used (Figure 2), which may be accentuated to identify diastole is to palpate the ca- Abnormal cardiac structure with the patient sitting upright and lean- Aortic regurgitation rotid artery pulse during auscultation; 9 ing forward. In some cases, S2 can be Mitral stenosis S is synchronous with the carotid ar- Pulmonic regurgitation 1 obscured by the murmur. Most AR mur- Tricuspid stenosis tery pulsation while S2 follows the pulse. murs are high pitched and are best heard Atrial myxoma A diastolic murmur is a diastolic sound Ventricular septal defect* with the diaphragm of the stethoscope Atrial septal defect* longer than a heart sound. Examiners placed firmly on the chest wall. Some AR Mitral regurgitation* should describe the grade, location of murmurs are low pitched and are bet- Normal cardiac structure, increased flow IGURE Renal failure with volume overload maximal intensity (F 1), timing ter heard with the bell of the stetho- Thyrotoxicosis (FIGURE 2), duration, pitch, and radia- scope placed lightly on the chest wall. Anemia tion of the murmur. Sepsis For example, the AR murmur associ- The Levine grading system,7 with Diastolic murmurs are caused by abnormally increased ated with endocarditis and a fenes- * 8 diastolic flow across the mitral or tricuspid valves. slight modifications, was developed for trated aortic valve can be low pitched. The examiner should apply the stetho- Figure 1. Typical Location of Abnormal Diastolic Murmurs scope to the chest wall in the third or

Figure 2. Selected Features of Diastolic Murmors

1 S S 2 2 1

S2 OS S1

Diastolic murmurs are classified based on the time of 15 3 onset of the murmur. An early diastolic murmur begins with the second heart sound (S2). Top, Early diastolic murmurs typically decrease in intensity (decrescendo) and disappear before the first heart sound (S1). In some cases, an early diastolic murmur can con- There are 3 important areas to auscultate for diastolic murmurs. Area 1 is the second and third intercostal spaces tinue through diastole. Bottom, A mid-diastolic mur- at the right-sternal border. Area 2 is the second and fourth intercostal spaces at the left-sternal border. Aortic mur begins clearly after S2 (in mitral stenosis, classi- regurgitation murmurs may be heard in both areas 1 and 2. If the murmur is loudest in area 1, then the un- cally after an opening snap [OS]). A late diastolic (or derlying cause of aortic regurgitation may be an ascending aortic aneurysm or aortic dissection. Pulmonic re- presystolic) murmur begins in the interval immedi- gurgitation murmurs are loudest in the superior part of area 2, and may radiate downward. The murmur of ately before S1. In mitral stenosis, the mid-diastolic mur- mitral stenosis and the Flint murmur of aortic regurgitation are best heard at the apex (area 3). mur may merge with the late systolic murmur.

Downloaded from www.jama.com at University of Florida, on May 24, 2005 AORTIC REGURGITATION fourth intercostal space at the left ster- Mitral stenosis is associated with of diastolic murmurs. There is no point nal border and listen while the patient a mid-diastolic decrescendo low- in doing maneuvers if a loud AR mur- stops his/her breath at end of expira- frequency rumble, which, if the patient mur has been detected during routine tion. The patient should not hold his/ is in sinus rhythm, may be followed by auscultation. However, if the clinician her breath because he/she may inadvert- late-diastolic (presystolic) crescendo that is unsure about the presence of a faint ently do a Valsalva maneuver. If the ends with the mitral component of S1 diastolic murmur, then a maneuver that murmur is louder at the second to third (Figure 2). It is best heard using the bell increases murmur intensity may clarify right intercostal space, the underlying of the stethoscope placed at the apex the situation. If the clinician has a height- cause of AR may be an ascending aortic soon after moving the patient into the ened suspicion for AR (eg, after hear- aneurysm or aortic dissection.10 left-lateral decubitus position. This will ing an aortic ejection sound), or if ex- Aortic regurgitation also may be as- both bring the left ventricle closer to the amining conditions are not optimal, then sociated with a systolic murmur,11 chest wall and serve as a form of exer- a maneuver to augment murmur inten- which results from the flow of an ab- cise, which will increase flow across the sity might bring out an otherwise inau- normally large volume of blood through mitral valve and, therefore, increase the dible murmur. Finally, if the clinician a nonstenotic aortic valve or a bicus- intensity of the murmur.10 The mur- is unsure whether a murmur repre- pid aortic valve. The murmur is an early mur of mitral stenosis may be inaudible sents PR or AR, then these maneuvers peaking, crescendo-decrescendo sys- in patients with low cardiac output. may help distinguish between the 2. In tolic sound that is best heard with the The S1 may be increased in intensity this latter situation, the clinician should 14 diaphragm of the stethoscope applied in mitral stenosis. A normal S1 is best listen where the murmur is just barely to the second-right intercostal space. appreciated near the apex where it should audible, so that it is easy to detect a de- The Flint murmur is a low-pitched be louder than S2. The S1 is normally crease or increase in murmur intensity late-diastolic apical murmur, which is as- softer than S2 in the second-right and sec- during the maneuver. sociated with AR. The murmur is likely ond-left intercostal spaces adjacent to the Quiet inspiration increases venous re- produced when the regurgitant jet of sternum. If S1 is as loud as or louder than turn and is intended to augment right- blood collides with the left ventricular the S2 in these areas, then the S1 is in- sided heart murmurs such as PR. To de- endocardium.12 The murmur may have creased in intensity. termine the effect of inspiration on the a mid-diastolic component, but the origi- An OS is a high-frequency early dias- intensity of the murmur, the exam- nal description by Flint referred only to tolic sound that is associated with the op- iner should listen during quiet inspi- “presystolic blubbering.”13 It is best heard ening of a stenotic mitral valve. It oc- ration rather than asking the patient to with the patient in the left-lateral decu- curs 50 to 100 milliseconds after the breathe deeply, as the murmur may be bitus position using the bell of the stetho- aortic valve component (A2)ofS2and is obscured by breath sounds. scope. Differentiating the Flint mur- best heard in the area from the left ster- Transient arterial occlusion primar- mur from the murmur of mitral stenosis nal border to the apex. Much like the ily increases systemic arterial resistance can be difficult. The murmur of mitral murmur of mitral stenosis, it may be ac- that results in an intensification of left- stenosis is primarily mid-diastolic (pos- centuated by listening to the patient in sided regurgitant lesions such as AR and sibly with a late diastolic component) and the left-lateral decubitus position shortly may help distinguish the murmur from may be associated with an opening snap after the patient has performed exer- that of PR. To perform this maneuver, 14 (OS) and a loud S1 (Figure 2). cise. The A2-OS interval shortens with in- sphygmomanometers are placed around The typical murmur of pulmonic re- creasing severity of mitral stenosis. The both of the patient’s arms and are inflated gurgitation (PR) is an early diastolic de- OS may be absent in the case of a heavily to 20 to 40 mm Hg above the previously crescendo murmur heard best in the sec- calcified immobile mitral valve. It is of- recorded systolic blood pressure. Any ond-left intercostal space at the sternal ten difficult to differentiate an OS from changes in murmur intensity are noted 16 border. The murmur may radiate to the the P2 of S2. The OS usually decreases 20 seconds after cuff inflation. third- and fourth-left intercostal spaces, in intensity with inspiration and S2-OS and may increase during quiet inspira- interval widens on standing. Con- Peripheral Hemodynamic Signs tion. If there is splitting of S2, the astute versely, P2 becomes louder with inspi- There are a variety of peripheral hemo- examiner may note that the murmur be- ration and the A2-P2 interval remains the dynamic signs traditionally associated gins after the pulmonic valve compo- same or narrows with standing.14 Addi- with AR. Some of these signs have been nent (P2)ofS2 rather than the aortic tionally, P2 is not expected to be heard adequately evaluated, including de Mus- component. The murmur of PR may be at the apex unless the patient has pul- set head bobbing sign,17 a wide pulse lower pitched than the murmur of AR, monary hypertension. pressure,18 the brachial-popliteal pulse unless pulmonary hypertension is pres- gradient (Hill sign19), Duroziez fem- ent. A right-sided Flint murmur can be Maneuvers oral murmur,17 the femoral pistol shot heard, particularly in patients with pul- Selective use of maneuvers can en- murmur,14 and Corrigan water ham- monary hypertension. hance the detection and interpretation mer pulse.20 The de Musset head bob-

Downloaded from www.jama.com at University of Florida, on May 24, 2005 AORTIC REGURGITATION bing sign consists of a forward shaking retrograde blood flow. If, however, the evant studies for additional informa- of the head with every heartbeat; it is best bruit becomes louder with proximal tion. Studies were excluded if they were observed in patients who are sitting.17 pressure (and softer with distal pres- review articles, involved patients Pulse pressure refers to the differ- sure), then this sign should not be used younger than 18 years, were small (ie, ence between systolic and diastolic as evidence of AR but may indicate the Ͻ20 participants), involved prosthetic blood pressures. A widened pulse pres- presence of a high-flow state such as re- heart valves, if no clinical examination sure may be defined as greater than 50 nal failure with volume overload.23 was performed or reported, or if there mm Hg.21 Other definitions include a Femoral pistol shot sounds are elic- was no acceptable reference standard pulse pressure greater than 50% of the ited by auscultating with the dia- (Doppler echocardiography or cardiac systolic pressure.18 Determination of the phragm of the stethoscope over the fem- catheterization). blood pressure has been described in oral arteries. A high-pitched pistol shot Studies were independently re- another article in this series.22 sound may be heard in AR. Corrigan viewed for methodological quality by the The brachial-popliteal pulse gradient water hammer pulse refers to an in- 2 authors and disagreements were re- (Hill sign) can be defined as a systolic creased volume and rate of rise of the solved by consensus. Quality grades were blood pressure in the lower extremities radial pulse when the wrist is elevated assigned using published guidelines.26 that is at least 20 mm Hg higher than that perpendicular to the body of a supine Grade A studies involve the indepen- in the arms.21 To determine a popliteal patient. The radial pulse should first be dent comparison of a sign or symptom blood pressure, an appropriately sized assessed while the patient is lying su- with a reference standard of diagnosis blood pressure cuff should be placed on pine with his/her arms resting at the among a large number of consecutive pa- the patient’s thigh14 with the artery sides. Sufficient pressure should be ap- tients suspected of having the target con- marker over the popliteal artery; the cuff plied to obliterate the pulse. While dition. Grade B studies meet the criteria should be inflated and the systolic pres- maintaining this pressure, the pa- for grade A studies but have a small num- sure can then be determined in the pop- tient’s arm should be elevated such that ber of patients. Grade C studies involve liteal fossa either by palpation, as judged it is perpendicular to the plane of the nonconsecutive patients, patients who by the point where the pulse reappears body. In AR, the pulse will become pal- are known to have the target condition as the cuff is deflated, or by ausculta- pable despite applying an equivalent and healthy individuals, nonindepen- tion, listening for Korotkoff sounds to ap- amount of pressure as when the arm dent comparisons between the sign or pear. Both the brachial and popliteal was at the patient’s side. symptom and the reference standard, or blood pressures should be measured Other peripheral hemodynamic signs, nonindependent comparison with a ref- while the patient is supine. The average such as Mayne sign (a decrease in di- erence standard of uncertain validity. of repeated readings should be used, es- astolic blood pressure of 15 mm Hg Grade C studies tend to overestimate the pecially in patients with irregular heart when the arm is held above the head accuracy of the sign or symptom. rates, such as atrial fibrillation. compared with when the arm is held We recreated contingency tables for Duroziez double intermittent fem- at the level of the heart),24 Quinke cap- all studies and determined the likeli- oral bruit is elicited by first gently com- illary pulsation, Muller pulsatile uvula, hood ratios (LRs) for the cardiac dis- pressing the femoral artery with the dia- and Rosenbach liver pulsation, have not ease of interest.27,28 When a cell from phragm of the stethoscope. This will been adequately evaluated for preci- the table included a value of zero, con- yield a systolic bruit in all patients. As sion or accuracy. fidence intervals (CIs) for the LRs were increasing pressure is applied to the dia- estimated using an iterative approach. phragm, an early diastolic bruit will be- METHODS We also sought information on the ex- come audible in patients with AR. To identify articles pertaining to the pre- amination for other causes of diastolic While listening to the diastolic bruit, cision and accuracy of the physical ex- murmurs, such as mitral stenosis or PR. the stethoscope should then be tilted amination for AR, we used standard Unfortunately, we found few studies of such that the distal rim (closest to the methods for conducting research over- sufficient methodological quality for patient’s feet) is compressing the fem- views.25 Our data collection strategy in- these conditions. This relative lack of oral artery. If the bruit becomes louder volved 3 steps and was deliberately broad information implies that methodologi- with this maneuver then the diastolic to reduce the possibility of overlook- cally sound studies are needed, but does bruit is due to the retrograde flow of ing important articles. First, we searched not imply that the clinical examina- blood toward the heart in AR. The MEDLINE for English-language ar- tion for these conditions is imprecise, stethoscope should then be tilted such ticles from 1966 through July 1997 inaccurate, or unimportant. that the proximal rim (closest to the pa- using a structured search strategy (avail- tient’s head) is compressing the fem- able on request from the authors). Sec- Precision of the Examination oral artery. If the diastolic bruit be- ond, we manually reviewed potentially Related to Diastolic Murmurs comes softer, this can be taken as relevant articles and their reference lists. Precision refers to agreement regarding supportive evidence of the presence of Third, we contacted the authors of rel- a particular clinical finding between dif-

Downloaded from www.jama.com at University of Florida, on May 24, 2005 AORTIC REGURGITATION ferent physicians (interobserver) or be- Table 2. Interobserver Reliability (Precision) for Detecting Diastolic Murmurs* tween multiple assessments by the same No. of No. of Simple physician (intraobserver). The preci- Finding Type of Examiner Examiners Patients ␬ Agreement, % sion of the clinical examination for di- Murmur absent vs present Cardiologists (tapes)31 5 100 0.51 79 astolic murmurs has been evaluated in Cardiologists29 2 32 ... 94 usual clinical situations by auscultating Noncardiologists29 3 32 ... 78 29,30 patients or in controlled nonclinical Intensity of murmur Not stated30† 5 25 . . . 92 circumstances by listening to recorded *Ellipses indicate data not available. 31 †Examiners used paper disks, 0.5 mm in thickness, that were progressively inserted between the chest wall and the audiotapes (TABLE 2). stethoscope until the murmur became inaudible. The total thickness of the disks used was used as the measure of There have been 4 studies that address intensity. For example, if a murmur was inaudible after inserting 3 disks, then this was a 1.5-mm murmur. theinterobserverprecisionofcardiacaus- cultation to detect diastolic murmurs lute height of the regurgitant jet or the (95% CI, 1.6-14.0) for distinguishing (Table2).Whilesimpleagreementishigh ratio of the height of the jet to the height severe AR from less severe AR, while a in these studies, the 1 study for which it of the left ventricular outflow tract on grade 2 murmur had an LR of 1.1 (95% was possible to calculate agreement ad- Doppler echocardiography), which CI, 0.5-2.4), a grade 1 murmur had an justed for chance (␬) showed only mod- allowed us to calculate LRs for detect- LR of 0.0 (95% CI, 0.0-0.9), and ab- erate agreement. The experience of ob- ing both mild (or worse) and moder- sence of a diastolic murmur had an LR servers likely affects precision. The 1 ate (or worse) AR. of 0.0 (95% CI, 0.0-1.1).48 study29 that compared cardiologists with Cardiologists conducted the clinical Two grade C studies of the Flint mur- noncardiologists found a higher simple examinations in most studies. Too few mur and some peripheral hemody- agreement for cardiologists. studies, using few patients, allow for rea- namic findings are reported in Table 3. The interobserver agreement be- sonable estimates of the accuracy of non- Grade C studies tend to overestimate tween examiners on the intensity of cardiologists, although noncardiolo- diagnostic test accuracy. Despite this heart sounds is excellent (92%).30 In this gists are likely less adept at detecting the tendency, 1 study found that absence study, examiners progressively in- diastolic murmur of AR. Approxi- of a Flint murmur did not rule out AR serted 0.5-mm-thick paper disks be- mately 20% of residents and medical stu- (negative LR, 0.5-0.8).49 Another study tween the patient’s chest and the stetho- dents correctly identified the murmur of of patients with mild-to-severe AR only scope. The total thickness of the disks AR on high-fidelity digitized audio- found that a wide pulse pressure or was used as a measure of heart sound tapes,32 while 46% of internal medicine peripheral hemodynamic sign (Duro- intensity. Murmur intensity was also as- residents correctly identified an AR mur- ziez bruit, femoral pistol shots, and Cor- sessed using this technique (Table 2). mur using a patient simulator.33 rigan pulses) was not helpful for dis- The best-studied physical finding is tinguishing mild AR from moderate or The Bottom Line for Precision the typical early diastolic murmur of severe AR.21 The interobserver precision of cardi- AR.34-47 If an examiner does not hear a One small study (grade C) evaluated ologists examining for any diastolic typical AR murmur then the likeli- peripheral hemodynamic signs in pa- murmur is moderate using audiotapes hood that the patient has moderate or tients exclusively with proven AR of vary- (␬ = 0.51) and good in the clinical set- greater AR is significantly reduced ing severity defined by aortography or ting (simple agreement, 94%). Noncar- (negative LR, 0.1 for grade A studies); surgery. These studies provide an esti- diologists may be less precise than car- the likelihood of mild or greater AR is mate of the sensitivity of the peripheral diologists. The precision of examining also significantly reduced (negative LR, hemodynamic signs. The de Musset head for the intensity of murmurs and heart 0.2-0.3 for grade A studies).34,35 bobbing sign was seen in only 1 of 20 pa- sounds using a standardized series of If an examiner hears the typical AR tients (sensitivity, 5%) and Duroziez fem- paper disks to assess intensity is good murmur, the likelihood that the pa- oral bruit was observed in 8 of 12 pa- (simple agreement, 92%-96%). tient has moderate or greater AR is in- tients (sensitivity, 67%).17 creased (positive LR, 4.0-8.3 for grade Accuracy of the A studies); the likelihood of mild or THE BOTTOM LINE FOR AR Examination for AR greater AR is also significantly in- When a cardiologist hears the typical We consider Doppler echocardiogra- creased (positive LR, 8.8-32.0 for grade murmur of AR, the likelihood of mild phy and cardiac catheterization to be A studies).34,35 or greater AR is increased signifi- acceptable reference standards for AR The intensity of the murmur corre- cantly (2 grade A studies). The ab- (TABLE 3). In 1 study, the reference stan- lates with the severity of echocardio- sence of a typical diastolic murmur sig- dard was open-heart surgery. Some graphic AR. Desjardins et al48 studied nificantly reduces the likelihood of AR studies explicitly graded the severity of 40 patients with echocardiographic AR, (2 grade A studies). Noncardiologists AR detected by the reference standard including 17 with severe AR. A grade may be less proficient than cardiolo- (for example, on the basis of the abso- 3 diastolic murmur had an LR of 4.5 gists at detecting the murmur of AR.

Downloaded from www.jama.com at University of Florida, on May 24, 2005 AORTIC REGURGITATION

Mitral Stenosis and PR of residents and medical students cor- murmur best audible in the second in- In 1 grade A study of 529 unselected rectly identified a mid-diastolic mur- tercostal space at the left-upper sternal nursing home residents (31 with mi- mur of mitral stenosis on a high- border, which may increase in intensity tral stenosis), a cardiologist detected a fidelity digitized audiotape,32 while 43% with quiet inspiration. All studies used mid-diastolic murmur in all cases of mi- of medical residents identified a mid- cardiologists as examiners and were of tral stenosis, with no false-positive or diastolic murmur of mitral stenosis us- poor methodologic quality (grade C). -negative examinations (W. A. Aronow, ing a patient simulator. In the latter When a cardiologist hears the mur- MD, written communication, 1997).50 study, only 21% identified the OS of mi- mur of PR, the likelihood of PR in- Only 1 patient had an audible OS. tral stenosis.33 creases (positive LR, 17 in both stud- Noncardiologists may be less profi- The only evaluated element of the ies), but absence of a PR murmur was cient at detecting the physical find- clinical examination for PR is the pres- not helpful for ruling out PR (LR, 0.9 ings of mitral stenosis. Less than 10% ence of a typical diastolic decrescendo in both studies).37,43

Table 3. Accuracy of the Physical Examination for Detecting Aortic Regurgitation* Positive Negative No. of Likelihood Likelihood Patients Ratio Ratio Quality Study Patient Population Reference Standard With AR (95% CI)† (95% CI)‡ Grade Typical Murmur With Severity of AR Specified Aronow and Kronzon,34 1989 Elderly patients Echocardiography (n = 450) A Mild or greater AR 131 32 (16-63) 0.2 (0.1-0.3) Moderate or greater AR 74 8.3 (6.2-11) 0.1 (0.0-0.2) Grayburn et al,35 1986 Referred for catheterization Catheterization (n = 106) A Mild or greater AR 82 8.8 (2.8-32) 0.3 (0.2-0.4) Moderate or greater AR 57 4.0 (2.5-6.9) 0.1 (0.1-0.3) Roldan et al,36 1996 Asymptomatic connective Echocardiography (n = 143) C tissue disease and Mild or greater AR 10 80 (14-470) 0.4 (0.2-0.7) controls Moderate or greater AR 5 69 (18-270) 0.0 (0.0-0.6) Rahko,37 1989 Referred for echocardiogram Echocardiography (n = 403) C Mild or greater AR 134 27 (13-60) 0.4 (0.3-0.5) Moderate or greater AR 82 12 (8.1-19) 0.2 (0.1-0.3) Cohn et al,38 1967 Mitral valve repair Open-heart surgery (n = 156) C Mild or greater AR 50 5.2 (3.3-8.4) 0.3 (0.2-0.4) Moderate or greater AR 37 3.9 (2.6-5.7) 0.2 (0.1-0.4) Meyers et al,39 1982 Referred for aortography Catheterization (n = 75) C Mild or greater AR 66 3.3 (1.3-12) 0.4 (0.2-0.7) Moderate or greater AR 39 1.6 (1.2-2.4) 0.4 (0.2-0.7) Dittmann et al,40 1987 Valvular heart disease Catheterization (n = 55) C§ Mild or greater AR 42 16 (2.1-155) 0.4 (0.3-0.6) Severe AR 19 3.6 (2.1-6.6) 0.1 (0.0-0.4) Meyers et al,41 1985 Valvular heart disease Catheterization (n = 20) C Mild or greater AR 11 9.8 (1.3-96) 0.5 (0.2-0.9) Moderate or greater AR 3 5.7 (1.4-14) 0.0 (0.0-0.9) Linhart,42 1971 Mitral stenosis Catheterization (n = 28) C Mild or greater AR 11 6.2 (1.9-23) 0.3 (0.1-0.7) Moderate or greater AR 7 7.0 (2.5-20) 0.0 (0.0-1.3) Typical Murmur Without AR Severity Specified (May Include Trivial AR) Come et al,43 1986 Mitral valve prolapse, plus Echocardiography (n = 165) 7 90 (8-982) 0.7 (0.4-0.9) C patients with systolic flow murmurs Nienaber et al,44 1993 Clinically suspected aortic Echocardiography (n = 110) 32 33 (9.4-120) 0.2 (0.1-0.3) C࿣ dissection Ward et al,45 1977 Clinically suspected aortic Catheterization (n = 65) 49 13 (2.9-75) 0.2 (0.1-0.3) C࿣ dissection Esper,46 1982 AR and other heart disease Echocardiography (n = 43) 24 12 (2.4-67) 0.4 (0.3-0.7) C Saal et al,47 1985 Mitral stenosis Catheterization (n = 45) 35 8.0 (1.9-45) 0.2 (0.1-0.4) C

Downloaded from www.jama.com at University of Florida, on May 24, 2005 AORTIC REGURGITATION

The Bottom Line volume overload, anemia, and hyper- study does not indicate how many of for Mitral Stenosis and PR tension.51 These murmurs typically dis- these patients had audible diastolic mur- Presence of a mid-diastolic murmur sig- appear after the treatment of volume murs. Audible diastolic murmurs may be nificantly increases the likelihood of mi- overload, as was demonstrated in 2 small relatively uncommon findings in asymp- tral stenosis, while absence of a mid- studies (grade C).51,52 These murmurs are tomatic persons. In 1 study only 1% diastolic murmur significantly reduces probably due to transient pulmonary hy- (n = 103) of elderly asymptomatic nurs- the likelihood of mitral stenosis (1 grade pertension and dilation of the pulmo- ing home residents had an audible dias- A study). When a cardiologist hears a nary artery root, leading to PR.52 tolic murmur.54 typical PR murmur, the likelihood of Despite the lack of evaluative data, PR increases significantly. Absence of The Bottom Line we think that a prudent clinician will a typical murmur does not alter the like- for Diastolic Murmurs examine for AR in most clinical set- lihood of PR (2 grade C studies). Non- in Patients With Renal Failure tings. Aortic regurgitation is a serious cardiologists may be less proficient at Although there is an insufficient cardiac abnormality, which may be detecting the mid-diastolic murmur of amount of data on which to make rig- caused by underlying disorders and may mitral stenosis. orous recommendations, if an early di- be asymptomatic. The clinician’s sus- astolic murmur is heard in a dialysis pa- picion for AR may be heightened by evi- Diastolic Murmurs in Patients tient with volume overload, the patient dence of systemic disease, such as an- With Renal Failure should be reexamined after treatment kylosing spondylitis, a peripheral Diastolic murmurs due to abnormal flow because the murmur may disappear. hemodynamic finding (although these states, rather than abnormal cardiac are by no means indicative of underly- structure, may be associated with a va- When to Examine for AR ing AR), or an abnormality detected riety of conditions. Renal failure with There are no evaluative data on which during routine auscultation (such as an volume overload is the only abnormal to base a recommendation regarding aortic ejection sound). Other findings flow state associated with diastolic mur- when to examine for AR. Undetected AR may suggest different cardiac abnor- murs that has been evaluated. may be common in elderly persons: 13% malities associated with diastolic mur- Up to 9% of patients with end-stage re- (n = 552) of asymptomatic elderly Finn- murs, such as evidence of pulmonary nal disease have diastolic murmurs, par- ish persons had moderate or severe echo- hypertension (for PR), a wide-fixed split 53 ticularly when these patients also have cardiographic AR. Unfortunately, that S2 (for atrial-septal defect), or a holo-

Table 3. Accuracy of the Physical Examination for Detecting Aortic Regurgitation* (cont) Positive Negative No. of Likelihood Likelihood Patients Ratio Ratio Quality Study Patient Population Reference Standard With AR (95% CI)† (95% CI)‡ Grade Maneuver With transient arterial occlusion Patients with AR, mitral Catheterization or 10 8.4 (1.3-81) 0.3 (0.1-0.8) C murmur increases stenosis, and pulmonic echocardiography in intensity16 regurgitation (n = 16) Associated Physical Findings Flint murmur 49 Isolated AR and controls Echocardiography (n = 36) C Mild or greater AR 28 4 (0.5-40) 0.8 (0.6-1.3) Moderate or greater AR 13 25 (2.8-243) 0.5 (0.2-0.7) Any systolic murmur 49 Isolated AR and controls Echocardiography (n = 36) C Mild or greater AR 28 1.3 (0.9-2.7) 0.5 (0.2-1.6) Moderate or greater AR 13 1.5 (1.0-2.1) 0.0 (0.0-1.0) Popliteal-brachial gradient Mild to severe AR Catheterization (n = 33) C Ͼ20 mm Hg21 Moderate or greater AR 28 8.2 (1.5-78) 0.2 (0.1-0.5) Peripheral hemodynamic signs21¶ Mild to severe AR Catheterization (n = 34) C Moderate or greater AR 28 2.1 (0.3-22) 0.8 (0.7-1.7) Pulse pressure Ͼ50 mm Hg21 Mild to severe AR Catheterization (n = 33) C Moderate or greater AR 28 1.0 (0.7-2.2) 0.9 (0.2-5.5) *AR indicates aortic regurgitation; CI, confidence interval. †The applicable likelihood ratio when the finding is present. ‡The applicable likelihood ratio when the finding is absent. §Grade B study except catheterization results not interpreted independently of clinical findings. ࿣Grade B study except echocardiogram not interpreted independently of clinical findings. ¶Included Duroziez bruit, femoral pistol shots, and Corrigan pulses.

Downloaded from www.jama.com at University of Florida, on May 24, 2005 AORTIC REGURGITATION systolic apical murmur (for mitral re- SCENARIO RESOLUTION quiet room with a comfortable and co- gurgitation). Your patient, who will be undergoing operative patient. You can advise the sclerotherapy for esophageal varices, primary care physician that AR is un- Recommendations has a wide pulse pressure but no typi- likely and echocardiography is not nec- for Further Research cal early diastolic murmur. The likeli- essary. Most studies used cardiologists to con- hood of mild or moderate AR is signifi- duct clinical examinations. There are cantly reduced by the absence of a Funding/Support: This work was funded in part by some data that suggest that noncardiolo- typical early diastolic murmur (nega- grant 98-01R from Physicians Services Incorporated Foundation. gists may be less accurate then cardiolo- tive LR, 0.1-0.3; 2 grade A studies). You Disclaimer: The views expressed in this article are those gists, so studies evaluating techniques perform transient arterial occlusion and of the authors and do not reflect those of any spon- soring or supporting agency. to improve the skills of noncardiologists no diastolic murmur appears, which en- Acknowledgment: We thank Eugene Oddone, MD, are needed. There are also no studies de- hances your confidence (negative LR, MHS, for his helpful comments on earlier drafts of this article, and W. S. Aronow, MD, and C. A. Nienaber, finingtheoptimalexaminationtechnique 0.3). You are confident in your assess- MD, for providing additional data and methodologi- for detecting the AR murmur. ment because it was conducted in a cal information about their studies.

REFERENCES 1. Scognamiglio R, Rahimtoola SH, Fasoli G, Nistri S, The clinical evaluation of aortic regurgitation. Arch In- wald E. Preoperative assessment of aortic regurgita- Volta SD. Nifedipine in asymptomatic patients with tern Med. 1965;116:357-365. tion in patients with mitral valve disease. Am J Car- severe aortic regurgitation and normal left ventricu- 22. Reeves RA. Does this patient have hyperten- diol. 1967;19:177-182. lar function. N Engl J Med. 1994;331:689-694. sion? JAMA. 1995;273:1211-1218. 39. Meyers DG, Sagar KB, Ingram RF, Paulsen WJH, 2. Dajani AS, Taubert KA, Wilson W, et al. Prevention 23. Blumgart HL, Ernstene AC. Two mechanisms in Romhilt DW. Diagnosis of aortic insufficiency. South Med of bacterial endocarditis. JAMA. 1997;211:1794-1801. the production of Duroziez’s sign. JAMA. 1933;100: J. 1982;75:1192-1194. 3. Ballard DJ, Khandeira BK, Tajik AJ, Seward JB, Weber 173-177. 40. Dittmann H, Karsch KR, Seipel L. Diagnosis and VP, Melton LJ. Population-based study of echocardi- 24. Abbas F, Sapira JD. Mayne’s sign is not patho- quantification of aortic regurgitation by pulsed dop- ography. Int J Technol Assess Health Care. 1989;5: gnomonic of aortic insufficiency. South Med J. 1987; pler echocardiography in patients with mitral valve dis- 249-261. 80:1051-1052. ease. Eur Heart J. 1987;8(suppl C):53-57. 4. Etchells E, Bell C, Robb K. Does this patient have 25. L’Abbe KA, Detsky AS, O’Rourke K. Meta- 41. Meyers DG, Olson TS, Hansen DA. Ausculta- an abnormal systolic murmur? JAMA. 1997;277: analysis in clinical research. Ann Intern Med. 1987; tion, M-mode echocardiography and pulsed doppler 564-571. 107:224-233. echocardiography compared with angiography for di- 5. Rushmer RF, Sparkman DR, Polley RFL, et al. Vari- 26. Holleman DR, Simel DL. Does the clinical exami- agnosis of chronic aortic regurgitation. Am J Cardiol. ability in detection and interpretation of heart mur- nation predict airflow limitation? JAMA. 1995;273: 1985;56:811-812. murs. AJDC. 1952;83:740-754. 313-319. 42. Linhart JW. Aortic regurgitation. Ann Thorac Surg. 6. Feinstein AR, Di Massa R. The unheard diastolic 27. Koopman PAR. Confidence intervals for the ra- 1971;11:27-37. murmur in acute rheumatic fever. N Engl J Med. 1959; tio of two binomial proportions. Biometrics. 1984;40: 43. Come PC, Riley MF, Carl LV, Nakao S. Pulsed dop- 260:1331-1333. 513-517. pler echocardiographic evaluation of valvular regur- 7. Freeman AR, Levine SA. The clinical significance 28. Centor RM, Keightley J. The Two by Two Ana- gitation in patients with mitral valve prolpase. JAm of the systolic murmur: a study of 1000 consecutive lyzer. v1.0. 1992. Coll Cardiol. 1986;8:1355-1364. cases. Ann Intern Med. 1933;6:1371-1385. 29. Raftery EB, Holland WW. Examination of the heart. 44. Nienaber CA, von Kodolitsch Y, Nicolas V, et al. The 8. Constant J, Lippschultz EJ. Diagramming and grad- Am J Epidemiol. 1967;85:438-444. diagnosis of thoracic aortic dissection by noninvasive im- ing heart sounds and murmurs. Am Heart J. 1965;70: 30. Sharma RS. Measurement of intensity of heart aging procedures. N Engl J Med. 1993;328:1-9. 322-326. sounds and murmurs. Lancet. 1981;2:612. 45. Ward JM, Baker CW, Rubenstein SA, Johnston 9. Constant J. Bedside Cardiology. 3rd ed. Boston, 31. Dubrow RJ, Calatayud JB, Abraham S, Caceres CA. SL. Detection of aortic insufficiency by pulse doppler Mass: Little Brown & Co Inc; 1985. A study of physician variation in heart-sound inter- echocardiography. J Clin Ultrasound. 1977;5:5-10. 10. Harvey WP. Cardiac pearls. Dis Mon. 1994;40: pretation. Med Ann Dist Columbia. 1964;33:305- 46. Esper RJ. Detection of mild aortic regurgitation 43-113. 308, 355. by range-gated pulsed doppler echocardiography. Am 11. Shaver JA. Cardiac auscultation. Curr Probl Car- 32. Mangione S, Nieman LZ. Cardiac auscultatory skills J Cardiol. 1982;50:1037-1043. diol. 1995;20:445-530. of internal medicine and family practice trainees. JAMA. 47. Saal AK, Gross BW, Franklin DW, Pearlman AS. 12. Landzberg JS, Pflugfelder PW, Cassidy MM, Schiller 1997;278:717-722. Noninvasive detection of aortic insufficiency in pa- NB, Higgins CB, Cheitlin MD. Etiology of the Austin Flint 33. St Clair EW, Oddone EZ, Waugh RA, Corey GR, tients with mitral stenosis by pulsed doppler echocar- murmur. J Am Coll Cardiol. 1992;20:408-413. Feussner JR. Assessing housestaff diagnostic skills us- diography. J Am Coll Cardiol. 1985;5:176-181. 13. Flint A. On cardiac murmurs. Am J Med Sci. 1862; ing a cardiology patient simulator. Ann Intern Med. 48. Desjardins VA, Enriquez-Sarano M, Tajiik AJ, Bai- 44:29-54. 1992;117:751-756. ley KR, Seward JB. Intensity of murmurs correlates with 14. Sapira JD. The Art and Science of Bedside Diag- 34. Aronow WS, Kronzon I. Correlation of preva- severity of valvular regurgitation. Am J Med. 1996; nosis. Baltimore, Md: Urban & Schwarzenberg; 1990. lence and severity of aortic regurgitation detected by 100:149-156. 15. Soffer A, Feinstein A, Rosener S, et al. Glossary pulsed doppler echocardiography with the murmur of 49. Lee D, Chen CH, Hsu TL, Chiang CE, Wang SP, of cardiologic terms related to physical diagnosis and aortic regurgitation in elderly patients in a long-term Chang MS. Reappraisal of cardiac murmurs related to history. Am J Cardiol. 1967;20:285-286. health care facility. Am J Cardiol. 1989;63:128-129. aortic regurgitation. Chin Med J (Engl). 1995;56:152. 16. Lembo NJ, Dell’Italia LJ, Crawford MH, O’Rourke 35. Grayburn PA, Smith MD, Handshoe R, Fried- 50. Aronow WS, Schwartz KS, Koenigsberg M. Cor- RA. Diagnosis of left-sided regurgitant murmurs by tran- man BJ, DeMarie AN. Detection of aortic insuffi- relation of murmurs of mitral stenosis and mitral regur- sient arterial occlusion: a new maneuver using blood ciency by standard echocardiography, pulsed dop- gitation with presence or absence of mitral annular cal- pressure cuffs. Ann Intern Med. 1986;105:368-370. pler echocardiography, and auscultation. Ann Intern cium in person older than 62 years in a long-term health 17. Sapira JD. Quincke, de Musset, Duroziez, and Hill. Med. 1986;104:599-605. care facility. Am J Cardiol. 1987;59:181-182. South Med J. 1981;74:459-467. 36. Roldan CA, Shively BK, Crawford MH. Value of 51. Alexander WD, Polak A. Early diastolic murmurs 18. Hegglin R, Scheu H, Rothlin M. Aortic insuffi- the cardiovascular physical examination for detect- in end-stage renal failure. Br Heart J. 1977;39:900. ciency. Circulation. 1968;37(suppl V):V77-V92. ing valvular heart disease in asymptomatic subjects. 52. Perez JE, Smith CA, Meltzer VN. Pulmonic valve 19. Hill L. The measurement of systolic blood pres- Am J Cardiol. 1996;77:1327-1331. insufficiency. Ann Intern Med. 1985;103:497-502. sure in man. Heart. 1909;1:73-82. 37. Rahko PS. Prevalence of regurgitant murmurs in 53. Lindroos M, Kupari M, Heikkila J, Tilvis R. Preva- 20. Warnes CA, Harris PC, Fritts HW. Effect of el- patients with valvular regurgitation detected by dop- lence of aortic valve abnormalities in the elderly. JAm evating the wrist on the radial pulse in aortic regur- pler echocardiography. Ann Intern Med. 1989;111: Coll Cardiol. 1993;21:1220-1225. gitation. Am J Cardiol. 1983;51:1551-1553. 466-472. 54. Bethel CS. Heart sounds in the aged. Am J Car- 21. Frank MJ, Casanegra P, Migliori AJ, Levinson GE. 38. Cohn LH, Mason DT, Ross J, Morrow AG, Braun- diol. 1963;11:763-767.

Downloaded from www.jama.com at University of Florida, on May 24, 2005