British Heart Journal, 1978, 40, 190-193

Pericardial effusions in patients with end-stagerenal disease1JAY H. KLEIMAN, JORGE MOTTA, ELIZABETH LONDON,J. PHILLIP PENNELL, AND RICHARD L. POPP2

From the Divisions of Cardiology and Nephrology, Stanford University School of Medicine,Stanford, California 94305, USA

SUMMARY Echocardiography has greatly increased the accurate recognition of pericardial effusion.Echocardiograms were performed prospectively on the total group of 35 stable asymptomatic patientson chronic haemodialysis to determine the incidence of pericardial effusion. Effusions were shown in1 1 per cent (4/35); only 6 per cent (2/35) were estimated as greater than 100 ml. For comparison, recordswere reviewed retrospectively from 41 haemodialysis patients referred during a 27-month period forechocardiographic assessment of suspected pericardial effusion. These 41 patients came from a totalgroup of 108 patients treated with chronic dialysis over this interval. Of 41 examined, 21 (510%) or 21 of108 (19%) of the population at risk had an effusion. Of 21 with echocardiographic effusions, 15 (71%),or 15 of41 (37%) ofthose with clinically suspected effusion, had more than 100 ml fluid. Gross ( > 100 ml)pericardial effusions are infrequent in stable, asymptomatic patients with end-stage renal disease. Whenclinical findings suggest pericardial disease, the echocardiographic demonstration of over 100 mlpericardial fluid is indicative of new effusion, rather than coincidental pre-existing effusion.

Pericarditis has been recognised as a clinical featureof uraemia for over 100 years, and frequently was aharbinger of death in the era before dialysis (Beaudryet al., 1966). More recently, recognition and treat-ment of uraemic pericarditis and its complicationshave allowed more patients to survive acute episodesand continue to be maintained on chronic haemo-dialysis.

Echocardiography has gained widespread use forthe non-invasive detection of pericardial effusion(Feigenbaum, 1972). Careful echocardiographicevaluation may reliably define effusions as small as20 ml (Horowitz et al., 1974). We were initiallyimpressed by the apparently high prevalence ofpericardial effusion among patients on maintenancehaemodialysis referred for echocardiography becauseclinical symptoms suggested pericardial disease.We considered the possibility that effusions mightbe present chronically among many patients withend-stage renal disease, perhaps as a result of

'This work was supported in part by an NIH Grant and waspresented in part at the 49th Scientific Session of theAmerican Heart Association, Miami, Florida, 1976.

2Dr Popp is a recipient of RCDA No. K04-HL-704-39.

Received for publication 19 July 1977

sustained volume overload. As long as such effusionswere small and did not cause circulatory embarrass-ment, they would remain clinically silent and un-detected. If this were the case, then the echocardio-graphic demonstration of effusion in patients onchronic dialysis would be of different clinical importthan if subclinical effusions were uncommon. Thecurrent study was undertaken to define more clearlythe significance of the echocardiographic demonstra-tion of pericardial effusion in these patients and toclarify the role of echocardiography in their manage-ment.

Methods

Thirty-five asymptomatic, stable patients com-prising the total group who had been dialysedchronically in our unit for at least 2 months (range 2to 72 months) underwent echocardiographic exam-ination during a 4-month study period (group A).None had any evidence of pericarditis during thestudy period. Of these 35 patients, 6 had hadprevious echocardiographic studies to rule out peri-cardial effusion at least 6 months before theirinclusion in this asymptomatic group. For com-parison, the medical records from a second group ofall 108 patients treated with chronic haemodialysis

190

on June 30, 2020 by guest. Protected by copyright.

http://heart.bmj.com

/B

r Heart J: first published as 10.1136/hrt.40.2.190 on 1 F

ebruary 1978. Dow

nloaded from

Pericardial effusions in patients with end-stage renal disease

during the 27-month period between June 1972 andSeptember 1975 were reviewed retrospectively. Ofthese 108 patients, 41 had been studied echocardio-graphically because clinical symptoms of fever, chestpain, friction rub, or persistent hypotension duringdialysis caused suspicion of pericarditis (group B).

Echocardiograms were performed with a Smith-Kline Instruments Ekoline Model 20A Ultrasono-

scope, which emits 1000 pulses/s. The transducerwas a 2-25 MHz, 13 mm active diameter model withan acoustic lens providing beam collimation to5 cm. Returning signals were recorded on a stripchart using either the Irex 101 or Honeywell 1856Visicorder. M-mode sector scanning was used in allcases. Lead II of the electrocardiogram was recordedsimultaneously with the echocardiogram. Theww w ~~~~~~~~~~~~~~~~~~~. . .. .. . h. ......~~a.5 . . 8 ....... <,~~~;pwHS <>iK --

EFS ..... a:

_~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~. .. ..M....

Fig. 1 Echocardiogram from an

asymptomatic stable patient. Aposterior echo-free space isobserved which extends into butnot through diastole. Thepericardium isflat relative to theepicardium. ECG, electrocardio-gram; IVS, interventricularseptum; En, endocardium;C, chord; Ep, epicardium; P,pericardium; EFS, echo-freespace.

ECG

,.. .........

MIJl Af-+E p-S ..~~~~~~~~~~~~~~~~~......;

. ^r ffl--.w ............ .... ....._-~~~~~~~..:cm

XF~~~~~~~~~~~~~.I S .....-

EFS

...~~~~~~~~~~~~~~~~~~..0..... :....::::..

... .. ....

:: .:~.'::::'.

Fig. 2 Classic echocardiographic pattern of pericardialeffusion, also from an asymptomatic stable patient. Aposterior echo-free space extends throughout systole anddiastole, and the pericardium remains flat relative to theepicardium. See Fig. I for key to abbreviations.

. ......

e.~

Fig. 3 Pericardial effusion with possible early pericardial

thickening. An additional linear echo (*) is present in

the echo-free space between the flat pericardium and the

epicardium. See Fig. I for key to abbreviations.

.......

.~~~~~~~~~~~~. .~~~~~~~~~~~... .. S. ..........*:::'~~~~~~~~~~~~~~.-:^.';2m.::....;-.. _.>,.. .:.:.:.. .....

ECGM.. M-

191

on June 30, 2020 by guest. Protected by copyright.

http://heart.bmj.com

/B

r Heart J: first published as 10.1136/hrt.40.2.190 on 1 F

ebruary 1978. Dow

nloaded from

12ay H. Kleiman, Jorge Motta, Elizabeth London, J. Phillip Pennell, and Richard L. Popp

patient's chest was raised approximately 30° fromthe horizontal, in the supine or shallow (300) leftlateral decubitus position. The transducer wasplaced in the third, fourth, or fifth intercostal spaceat the left sternal border. The chosen transducerposition permitted recording of the echoes from themitral valve leaflets while the transducer was per-pendicular to the chest wall in the sagittal plane.Inferolateral angulation of the transducer from thisposition allowed recording of the echoes from theventricular structures.A technically satisfactory study was defined as the

simultaneous recording of echoes from the posteriorleft ventricular pericardium, epicardium, endo-cardium, mitral chordae tendineae, interventricularseptum, and anterior heart wall. In order to locatethe strong pericardial echo, the damping control wasabruptly increased after the optimal gain settingsfor the ventricular structures had been achieved(Feigenbaum, 1972). Continuous recording of thisarea over several cardiac cycles at maximal dampingpermitted the clear observation of isolated peri-cardial movement. Slightly less damping oftenbrought out the epicardium and occasionally a smallseparation between the posterior epicardium andpericardium. As the damping was further reduced,this small separation often was obscured. Minimaldamping allowed identification of all ventricularstructures and verified the patterns of pericardial,epicardial, and endocardial echoes of a satisfactorystudy. The volume of pericardial fluid was estimatedas described by Horowitz et al. (1974).

In order to attempt to determine if myocardialcontractility was altered in uraemic patients at thetime that effusions were present, the fractionalshortening was measured in symptomatic andasymptomatic patients. The left ventricular internaldiameters at end-diastole (LVIDd) and end-systole(LVID5) were measured and used to calculatefractional shortening (FS), an index of circum-ferential myocardial contractility (McDonald et al.,1972; McDonald, 1976), which was defined as

LVIDd - LVIDs(%)FS LVIDd x 100 [1]

Normal values for fractional shortening in ourlaboratory are 28 to 41 per cent. Informed consentwas obtained from all asymptomatic patients.

Results

Of the 35 asymptomatic patients in group A, 4(11%) had pericardial effusions and only 2 (6%)had effusions estimated as more than 100 ml(Horowitz et al., 1974) (Fig. 1-3). This was indistinct contrast to the findings among the sympto-

Table Left ventricular fractional shortening in patientswith end-stage renal disease

No. Fractional shortening(%)

Asymptomatic stable 23 36-2 ±1-8

Symptomatic: 1 *

With effusion 20 30-6 2-3-NS j

No effusion 19 26-0 2-3

*p < 0 05 [normal = 28-41%1.NS = not significant.

matic patients in group B. Of 108 patients at riskduring the study period, 41 had echocardiogramsperformed because clinical symptoms suggestedpericardial disease and 51 per cent (21/41) had aneffusion. Six of these 41 patients had small effusionsestimated to be less than 100 ml, while 15 of 41patients (37%) had effusions greater than 100 ml.Some patients had signs we interpret as organisationof the pericardial fluid or pericardial thickening(Fig. 3).Records of excellent technical quality satisfactory

for calculation of fractional shortening wereobtained in 23 patients in group A and 39 patientsin group B (Table). Mean values for fractionalshortening were similar among symptomaticpatients with or without pericardial effusions, butthe values were less than those among asymptomaticpatients (Table). Eleven patients with effusions hadmultiple (2 to 7) echocardiographic examinationsperformed, either to follow the course of a singleeffusion or to assess several episodes of pericardialdisease. Improvement or deterioration in ventricularfunction as measured by fractional shortening didnot correlate with changes in the volume of effusionin this small series.

Discussion

Uraemic pericarditis continues to cause significantmorbidity and mortality among patients in haemo-dialysis programmes (Bailey et al., 1968; Comty etal., 1971; Baldwin and Edwards, 1976). Past studiesof clinically evident uraemic pericarditis haveindicated a prevalence of between 16 and 41 per centamong patients in chronic dialysis programmes.Many patients have developed pericarditis 3 monthsor longer after starting dialysis, in spite of tech-nically adequate and frequent dialysis (Beaudry etal., 1966; Bailey et al., 1968; Comty et al., 1971). Inthe past, only frank pericarditis or tamponade whichproduced signs such as chest pain, friction rub, orhypotension have been detected. The clinical

192

on June 30, 2020 by guest. Protected by copyright.

http://heart.bmj.com

/B

r Heart J: first published as 10.1136/hrt.40.2.190 on 1 F

ebruary 1978. Dow

nloaded from

Pericardial effusions in patients with end-stage renal disease

assessment of rises in the jugular venous pulse hasproved unreliable because obvious and sometimesrapid changes in intravascular volume are associatedwith dialysis. Several investigators have observedthat measurements of blood urea nitrogen orcreatinine do not predict the patients who willdevelop pericarditis (Comty et al., 1971).With the advent of echocardiography, a simple,

sensitive, and uniformly safe method to diagnosepericardial effusion has become available. We usedechocardiography to study prospectively a group ofasymptomatic stable patients undergoing chronichaemodialysis. This method allows reliable identi-fication of patterns associated with even smallamounts of pericardial fluid, but we did not attemptto quantify the volume of effusion when the echo-cardiographic pattern suggested only 20 ml to 100 mlof pericardial fluid. Unsuspected effusions of 100 mlor larger were infrequent (6 %). For comparison, theechocardiograms were reviewed from 41 patientswho had symptoms suggestive of pericardial diseaseat the time of study. In 51 per cent of these patients,pericardial effusions were shown, and in 37 per centthe effusions were estimated to be greater than100 ml. These observations imply that when clinicalfindings suggest pericardial disease, the echocardio-graphic demonstration of 100 ml or more of peri-cardial fluid is highly indicative of new effusion,rather than of coincidental pre-existing effusion. Webelieve such patients warrant careful and continuingclinical observation, as well as frequent echocardio-graphic follow-up.Necropsy studies of patients who have died as a

result of uraemic pericarditis have shown evidenceof an associated myocarditis, suggesting possibleimpairment of ventricular function (Langendorf andPirani, 1947; Gouley, 1940; Baldwin and Edwards,1976). Our current echocardiographic assessment ofleft ventricular function in small numbers ofpatients showed that fractional shortening was notuniformly depressed when pericardial effusionswere present, and that changes in left ventricularperformance did not parallel changes in the volumeof pericardial effusion measured echocardio-

graphically. The measurement of change in leftventricular dimensions is subject to error in thosepatients with sufficient pericardial effusion to allowthe heart to swing about within the pericardium, sothe sound beam strikes different parts of the walls atend-diastole and end-systole. It is interesting thatthe patients with symptoms but no effusion had thelowest values for fractional shortening. Additionalprospective studies of larger numbers of patientswith end-stage renal disease will be necessary toclarify these findings.

References

Bailey, G. L., Hampers, C. L., Hager, E. B., and Merrill, J. P.(1968). Uremic pericarditis: clinical features and manage-ment. Circulation, 38, 582-591.

Baldwin, J. J., and Edwards, J. E. (1976). Uremic pericarditisas a cause of cardiac tamponade. Circulation, 53, 896-901.

Beaudry, C., Nakamoto, S., and Kioff, W. J. (1966). Uremicpericarditis and cardiac tamponade in chronic renal failure.Annals of Internal Medicine, 64, 990-995.

Comty, C., Cohen, S., and Shapiro, F. L. (1971). Pericarditisin chronic uremia and its sequels. Annals of InternalMedicine, 75, 173-183.

Feigenbaum, H. (1972). Echocardiography, p. 163. Lea andFebiger, Philadelphia.

Gouley, B. A. (1940). Myocardial degeneration associatedwith uremia in advanced hypertensive disease and chronicglomerular nephritis. American J7ournal of the MedicalSciences, 200, 39-49.

Horowitz, M. S., Schultz, C. S., Stinson, E. B., Harrison,D. C., and Popp, R. L. (1974). Sensitivity and specificity ofechocardiographic diagnosis of pericardial effusion.Circulation, 50, 239-247.

Langendorf, R., and Pirani, C. L. (1947). The heart in uremia:an electrocardiographic and pathologic study. AmericanHeart Journal, 33, 282-307.

McDonald, I. G. (1976). Echocardiographic assessment of leftventricular function in aortic valve disease. Circulation, 53,860-864.

McDonald, I. G., Feigenbaum, H., and Chang, S. (1972).Analysis of left ventricular wall motion by reflected ultra-sound: application to assessment of myccardial function.Circulation, 46, 14-25.

Requests for reprints to Dr Richard L. Popp,Cardiology Division, Stanford University School ofMedicine, Stanford, California 94305, USA

193

on June 30, 2020 by guest. Protected by copyright.

http://heart.bmj.com

/B

r Heart J: first published as 10.1136/hrt.40.2.190 on 1 F

ebruary 1978. Dow

nloaded from