circulation 1972 selzer 878 90

Natural History of Mitral Stenosis:A Review

BY ARTHUR SELZER, MED., AND KEITH E. COHN, M.D.

T HE INTRODUCTION of cardiovascu-lar surgery has profoundly influenced

the natural history of many cardiac diseases.Availability of surgical means of benefitingserious heart disease-regardless of whetherthese means are curative of palliative-pro-vides a powerful temptation to take advantageof such an opportunity. It is evident, however,that in many instances medical therapy maycarry a much smaller risk than that of surgicalintervention. It is therefore crucial to considerthe natural history of a disease before adecision regarding surgery is made.The natural history of many diseases can no

longer be observed, since it would beimproper to withhold surgical help in certainadvanced stages. Consequently, natural his-tory often has to be extrapolated from olderstudies, which are incomplete by today'sstandards, and from observations of patientswho for various reasons were not subjected tosurgical therapy. In this review, we haveattempted to analyze critically the naturalhistory of mitral stenosis as it can bereconstructed today, and from these observa-tions have attempted to formulate a rationaltherapeutic approach to this disease.

EtiologyThe traditional view that diseases of left-

sided cardiac valves are late consequences ofrheumatic fever no longer holds true, for

From the Cardiology Unit, Presbyterian Hospitaland the Heart Research Institute, the Institute ofMedical Sciences, Pacific Medical Center, SanFrancisco, California.

Supported by Grants TI HE-5498 and HE-06311from the National Institutes of Health.

Address for reprints: Dr. Arthur Selzer, Director,Division of Cardiology, Presbyterian Hospital ofPacific Medical Center, P. 0. Box 7999, SanFrancisco, California 94120.

878

isolated lesions of the aortic valve and puremitral regurgitation, more often than not, arecaused by other etiologic factors. Mitralstenosis, alone or in combination with otherforms of valvular disease, is the only lesionattributed principally to rheumatic disease.The relationship between streptococcal in-

fection and rheumatic fever is well supportedby epidemiologic, clinical, and experimentalobservations.1 4 A recent piece of supportingevidence is by Goldstein et al.5 whoshowed that group A streptococci haveantigens which cross react with the structuralglycoprotein of the heart valves, thus directlylinking this organism with rheumatic valvulardisease. Rheumatic fever tends to be mimetic;hence, in a recurrent attack the host generallyremains free of carditis if he was initially freeof it and to have recurrent carditis when thiswas present initially.6 Of equal importance hasbeen the observation that permanent valvulardamage generally appears only in thosepatients with previous acute carditis; patientshaving arthritis but no carditis are usuallyspared chronic valvular damage.6 These find-ings have rendered support for prophylacticpenicillin therapy in some patients with priorrheumatic fever, although the value of thisprophylaxis in patients who had no evidenceof carditis remains debatable. After severalyears have elapsed from the last rheumaticattack, the probability of a recurrence dimin-ishes and the need for penicillin lessens.Involvement of the mitral valve represents

the most common of all locations of therheumatic process. Nevertheless, some puz-zling aspects have engendered an inquiry intothe exclusive role of the rheumatic process incausing mitral stenosis. A large proportion ofpatients with mitral stenosis-as many as halfin some series-have no history of rheumatic

Circulation, Volume XLV, April 1972

by guest on May 24, 2015http://circ.ahajournals.org/Downloaded from

MITRAL STENOSIS

fever. The usual and most plausible explana-tion given is that patients with mitral stenosiswho have no recollection of acute rheumaticfever suffered from subclinical rheumaticcarditis; rheumatic fever is particularly apt toescape attention if unassociated with polyarth-ritis. Burch and Colcolough, searching for analternate etiologic factor, brought out thepossibility of viral infection producing theanatomic changes leading to mitral stenosis.An exceedingly rare form of mitral stenosis iscongenital stenotic malformation of the mitralvalve; this is associated with serious circula-tory disturbances and is seldom seen beyondthe age of 3 years.8The association of mitral stenosis with

congenital defects of the atrial septum hasbeen widely recognized since the descriptionby Lutembacher. Mitral stenosis as a part ofthis syndrome is usually considered to berheumatic in origin.9 Edwards10 suggestedthat the interatrial communication may be aforamen ovale, stretched secondarily to leftatrial hypertension, as a result of mitralstenosis. Recently an interesting possibilitywas suggested by Okada et al.,1' whopostulated that mitral valve changes in theLutembacher syndrome are caused by anonrheumatic fibrosis of the mitral valve.representing a response to trauma fromaltered stresses and abnormal flow patternsthrough the mitral orifice. Each of thesemechanisms is probably capable of producingthe Lutembacher syndrome.

Clinical Evolution of Mitral StenosisThe course of mitral stenosis, from its

inception during acute carditis to the pointwhere surgical treatment becomes mand-atory,usually spans a period of several decades.Little factual information is available regard-ing the early stages of mitral stenosis. Duringthe initial attack of rheumatic carditis, themitral valve leaflets are commonly affected;however, such changes are incapable ofproducing actual narrowing of the valveorifice. Clinical signs of mitral regurgitationare found in these initial stages, although suchfindings do not necessarily indicate permanentCirculation, Volume XLV, April 1972

damage to the mitral valve, often beingcaused by temporary malfunction of themitral valve-papillary muscle-chordal appara-tus.

It is generally agreed that a time interval ofseveral years has to elapse between the initialattack of carditis and the time when definiteclinical evidence of mitral stenosis becomesapparent. The traditional view12 that this timeinterval varies from 2 to 8 years is probably anunderestimate. Bland and Jones,13 in their 20-year follow-up study of children with acuterheumatic fever, showed that nearly twothirds of those who had mitral stenosis at thetermination of the study did not show it at the10-year half point. Hence, it appears that inthe majority of cases the process of develop-ment of mitral stenosis takes longer than adecade.The majority of patients with fully devel-

oped mitral stenosis remain asymptomatic fora varying length of time. Thus, there is a"latent" period of mitral stenosis, which mightbe subdivided into two stages: first, the stageof formation of mitral stenosis, and second,the asymptomatic stage of fully developedmitral stenosis. Wood's14 series showed thatthe latent period lasted an average of 19years: the mean age for the attack of carditiswas 12 years and the age at the appearance ofsymptoms, 31 years. Wood14 also estimatedthat from the onset of symptoms to the stageof total disability, an average of 7 yearselapsed.The subject of clinical progression of mitral

stenosis was investigated in two longitudinalstudies. Rowe et al'5 followed a group of 250patients with pure mitral stenosis for 20 years,or until death. About half of the patients wereunder 30 years of age at the inception ofstudy; 52% were asymptomatic (class I). Atthe end of the study, 79% of the patients weredead, and 13% remained unchanged. After thefirst 10 years, 39% were dead. Of the patientswho were initially asymptomatic, 59% re-mained unchanged after 10 years, and 24% didso after 20 years. The other study'6 dealt withpatients who were already symptomatic(average age at inception of study was 41.5

879


SELZER, COHN

years). After 11 years, 11% remained un-changed, and 70% were dead; after 18 years,3% were unchanged, and 83% were dead. Theseobservations plainly show the potentiallyserious nature of mitral stenosis. When mitralvalvulotomy was introduced, it was clearlydemonstrated on the basis of the above figuresthat the course of patients with significantsymptoms was favorably influenced by theoperation.17 However, it is equally apparentfrom such figures that a proportion of patientswith mitral stenosis may remain clinicallyunchanged for many years, or, to put in otherterms, asymptomatic patients may remainasymptomatic indefinitely, while mildly symp-tomatic patients may remain unchanged for aslong as 20 years.Most patients with mitral stenosis tend to

develop symptoms in the fourth or fifthdecade of life. In about half, symptomsdevelop gradually, and, in the other half,abruptly-often precipitated by a complica-tion such as atrial fibrillation, in our experi-ence. When patients become symptomatic,some can still be controlled adequately bymedical management for long periods of time;others may deteriorate rapidly into uncontrol-lable disability such that surgical treatmentbecomes necessary.

During the past few years, important newinformation has unfolded showing that thecourse of mitral stenosis in areas with a lowerstandard of living differs greatly from thatsummarized above. In "depressed" areas,181-2mitral stenosis tends to progress rapidly andmay lead to serious disability early in life, sothat surgical treatment becomes necessary.Using the percentage of mitral valve opera-tions performed in patients under 20 years ofage as an index, the comparison of databetween various areas is presented in table 1.It is seen that economically deprived areas ofthis country may also produce an alteredepidemiologic pattern of rheumatic heartdisease: University of Kentucky MedicalCenter, which draws its clinical materialheavily from the population of the neighbor-ing Appalachia, has seven to 15 times higher

Table 1Mitral Valve Operations Performed in Patientsunder 20 Years of Age

Location % of operations

San Francisco* 0.5Philadelphia24 1.0Edinburgh25 1.0Italy'8 3.0Poland" 5.0Lexington, Kentuckyt 7.5Israel'" 8.0India22 27.0India20 34.0Iraq'1 40.0

*Our own series.tNoonan JA: Personal communication.

incidence of symptomatic juvenile mitralstenosis than most other centers.Complications in the Natural Historyof Mitral Stenosis

Atrial fibrillation is the most commoncomplication, perhaps even sequel, to mitralstenosis. The overall incidence of atrialfibrillation in this condition is estimated to beabout 40%.14 Atrial fibrillation often developsfirst in a paroxysmal form; later, it may appearin persistent form, but responds to antiar-rhythmic therapy. Eventually atrial fibrillationestablishes itself permanently in a therapy-resistant form. Atrial fibrillation is related toleft atrial enlargement and left atrial hyper-tension. Initially appearing as a functionaldisturbance in the electrophysiologic mecha-nism, perhaps as a response to stretching ofthe atrial musculature, it eventually is perpet-uated by organic changes, namely, a disinte-gration of the architecture of the atrialmuscle.26 Patients with very large left atriaand those who have had atrial fibrillation forlonger than 5 years are likely to have extensivedisruption of their atrial musculature (i.e.,atrial fibrosis and muscle atrophy) andtherefore are unlikely to respond to DC shockor antiarrhythmic therapy, even after asatisfactory surgical correction of the mitralstenosis.

Atrial fibrillation has a profound effect uponthe natural history of mitral stenosis. Hemo-dynamically, it causes lower resting cardiac


880


MITRAL STENOSIS

output at comparable ventricular rates.27 Theonset of atrial fibrillation is the most commonfactor bringing a previously asymptomaticpatient into a stage of disability. The rapidventricular rate of a suddenly appearing atrialfibrillation produces many emergencies inpatients with mitral stenosis. Furthermore,even after the ventricular rate is broughtunder control, patients often find themselvesin a lower effort capacity than prior to theonset of the arrhythmia, due to loss of theatrial "kick" mechanism.

Systemic emboli are among the mostdreaded complications of mitral stenosis.Their incidence is given as 9-14% of patientswith mitral stenosis, with 60-75% of thosehaving cerebral embolism.'4' 28-30 Systemicembolization occurs primarily in the presenceof atrial fibrillation. It is not known whetheratrial fibrillation is a sine qua non of thiscomplication, with those reported as havingembolism while in sinus rhythm possiblysuffering from an unrecognized paroxysm offibrillation, or whether mural thrombi mayactually develop in the contracting atrium.The disastrous consequences of cerebral

embolization are often compounded by thefact that patients with mitral stenosis who areasymptomatic, or even unaware of theirdisease, may be stricken. The occurrence ofsystemic emboli in patients who are notdisabled by mitral stenosis poses a significantdilemma regarding surgical therapy, especiallyif mitral stenosis is shown to be mild. It isevident that mitral valve operation does noteliminate the danger of embolism. Whether theincidence of embolization is reduced by valvot-omy has not been answered; studies pur-porting to show such beneficial effect lackstatistical design to make preoperative andpostoperative comparison valid.30 Severalstudies have indicated the favorable effect ofanticoagulant therapy upon the incidence ofrecurrent emboli; the evidence for this is atbest suggestive, for the experimental designof available studies and controls is not ade-quate for a conclusive answer."Myocardial factor" is considered by some

an additional cause of cardiac failure, un-Circulation, Volume XLV, April 1972

related to the mechanical effects of mitralstenosis.3' An impairment of left ventricularfunction, although detectable by sensitivehemodynamic techniques, is generally inconse-quential from the clinical standpoint, exceptin some elderly female patients with chronicatrial fibrillation, in whom obstruction at themitral valve is mild and clinical disability isrelated to a low cardiac output. The myocar-dial factor is often blamed for the failure ofpatients to show satisfactory improvementafter cardiac surgery.32 However, such inter-pretation should be questioned unless hemo-dynamic proof of nearly complete eliminationof mitral block is present. Patients whoundergo open-heart surgery occasionally de-velop myocardial damage sustained at thetime of perfusion, which should not beconfused with the "myocardial. factor" ofmitral stenosis.

Other complications of mitral stenosisshould be mentioned as affecting its naturalhistory: respiratory infections occur morefrequently in patients with mitral stenosis thanin the general population and may precipitatecardiac failure. Infectious endocarditis is veryrare in pure mitral stenosis,33 but its incidenceis significant in the patients who, in addition,have aortic or mitral regurgitation. Occasion-ally, patients with mitral stenosis developmassive pulmonary hemorrhage ("pulmonaryapoplexy") . Some reports indicate thatemergency operation upon the mitral valvemay be lifesaving in severe, exsanguinatingpulmonary hemorrhage.34Hemodynamic Evolution of Mitral Stenosis

Studies performed with the aid of cardiaccatheterization show in a most reliable mannerthe severity of mitral stenosis and its conse-quences upon the circulation. The importanceof hemodynamic evaluation of mitral stenosiscannot be overemphasized in patients inwhom surgical treatment is being contemplat-ed, for both the symptomatology and theclinical signs of mitral stenosis are oftenmisleading.The degree of physiologic stenosis of the

mitral valve can be assessed with the aid of

881


SELZER, COHN

the formula by Gorlin and Gorlin.45 In spite ofcertain oversimplifications and assumptions inthe application of the hydraulic formula (e.g.,constant flow through the valve), the calcula-tion of the mitral valve orifice gives reproduci-ble values, which check reasonably well withsurgical estimates of it. The accuracy of theGorlin formula depends upon the care withwhich measurement of cardiac output isperformed. The pressure gradient across thevalve is now most often measured directly,although the formula was originally describedwith the use of an assumed level of leftventricular diastolic pressure. Inasmuch as thevalve area is related to the square root of thepressure gradient, its accuracy and reproduci-bility are best in the presence of largegradients, i.e., higher degrees of stenosis, andleast in mild stenosis.

In the early stages of mitral stenosis,findings at cardiac catheterization may beentirely nornmal and there may be no pressuregradient, or one too small to be measured withavailable instrumentation. The normal atrio-ventricular filling occurs through a valveorifice of approximately 3 cm2. Mildestdegrees of valve obstruction (between 2 and 3cm2) are virtually unmeasurable. Mild mitralstenosis (a label usually applied to valve areasbetween 2 and 1.4 cm2) seldom producessymptoms. Moderate stenosis (in the rangebetween 1.4 and 0.9 cm2) and severe stenosis(less than 0.9 cm2) constitute the populationfrom which surgical candidates are selected.The most important consequence of mitral

stenosis is left atrial hypertension, which inturn elevates pressures in the pulmonaryvascular system and is largely responsible fordyspnea, the principal cause of patient'sdisability. Since mitral stenosis develops veryslowvly and gradually, pressure increments inthe pulmonary capillary system are small,permitting adaptive processes to take place.The most effective compensatory mechanismpreventing excessive fluid from being drivenout of the pulmonary capillaries under highhydrostatic pressure (pulmonary edema) isthe increased capacity of the lymphatic systemto drain excess fluid. Thus the degree of

dyspnea and the incidence of attacks ofpulmonary edema are much lower in mitralstenosis than in the suddenly appearing leftventricular failure with equivalent left atrialhypertension.The level of left atrial pressure is related to

the severity of mitral stenosis. However, it ismaterially modified by such factors as cardiacoutput and cardiac rate, and indirectly influ-enced by atrial size and compliance as well asby the circulating blood volume (hence, theinfluence of diuretics). Pronounced swings ofleft atrial and pulmonary capillary pressuremay occur in patients who happen to have arelatively high-output state, wherein, even withmild mitral stenosis, pulmonary edema maydevelop during exercise or tachycardia. On theother hand, some elderly individuals withchronic atrial fibrillation and low, relativelyfixed, cardiac output may have minimalsymptoms in the presence of significant mitralstenosis.As mitral stenosis progresses and reaches

the "severe" stage, some individuals developsecondary pulmonary arteriolar constriction,leading to severe pulmonary hypertension.14 3CAn index of this "reactive" pulmonary hyper-tension is the level of pulmonary vascular re-sistance, which can reach a level as highas five to 10 times the normal level. Thisactive" component is superimposed on "pas-

sive" pulmonary hypertension, i.e., that pro-duced as a direct effect of increased left atrialpressure. Pulmonary arterial pressures mavreach systemic levels in some cases.The hemodynamic consequence of such

severe pulmonary hypertension is an overloadupon the right ventricle, often producing rightventricular failure and tricuspid regurgitation.Inasmuch as low cardiac output limits theelevation of pressure in the left atrium and inthe pulmonary capillaries, pulmonary hyper-tension has been considered by some a"protective" mechanism, preventing pulmo-nary edema. Convincing evidence for this isnot available, and the "protective" effect ofhigh pulmonary vascular resistance may bemore a fallacy than a reality. Although suddenepisodes of frank pulmonary edema probably


882


MITRAL STENOSIS

occur more often in patients with normal oronly slightly elevated pulmonary arterialresistance, patients with very high resistancedo manifest high resting and exercise pulmo-nary capillary pressures and have disabilityfrom pulmonary congestion and edema.37Thus, they are in no way protected from mostof the ravages of severe mitral stenosis. Highpulmonary arterial resistance is reversible ifimprovement of left atrial pressure occursafter surgical treatment, in contrast to congen-ital heart disease, where organic and irreversi-ble pulmonary vascular disease is usuallyresponsible for the high pulmonary arterialresistance.38

Previous knowledge of the progression ofmitral stenosis was mainly based upon com-parison of hemodynamic findings in patientswith varying severity of mitral stenosis.Recently, we have undertaken a longitudinalstudy of patients with mitral stenosis, in orderto develop more insight into the modes ofhemodynamic progression of this disease.39Serial studies performed for periods of from 1to 10 years reveal that mitral stenosis pro-gresses in some patients at a rapid rate, whilein others it remains relatively stable. Progres-sion, defined as a narrowing of the mitralvalve orifice, is usually associated with aggra-vation or development of clinical disability. Inthe nonprogressive form, where the mitralvalve area remains stable, symptoms developless frequently and, when present, are usuallyrelated to complications, most commonlyatrial fibrillation. It is noteworthy that suchprogressive and nonprogressive forms havebeen demonstrated not only in patients whonever underwent surgical valvotomy, but alsoin those following such operations ("resteno-sis"').Pathologic Evolution of Mitral StenosisDuring acute rheumatic fever with carditis,

involvement of the mitral valve consists oftiny, translucent nodules located along the lineof closure of the valve, occasionally alsoinvolving subvalvular parts of the chordae.40Aschoff bodies are not usually encounteredupon the valve tissue. Microscopic sections ofCirculation, Volume XLV, April 1972

these nodules show largely nonspecific prolif-eration of fibroblasts and macrophages. Thesetranslucent vegetations later become opaqueand gray, and eventually more of the valveleaflet becomes thickened. Changes within thevalve structure involve deposition of fibrinupon the cusps with loss of the normalmorphology, hyalinization, and eventually thecovering of the leaflets with endothelium.4This process may lead to fusion of the valvecommissures. Brock42 postulated that theinitial point of fusion of the two leafletsoccurred at the "critical area of tendoninsertion," i.e., the point where the shortestand most direct chordae connect with thecusps. When fusion occurs at these points,portions of the cusps lateral to them areimmobilized, thereby facilitating more com-missural fusion.

Virtually no information is available regard-ing early progression of pathological changesin the mitral valve. When mitral stenosis isfully developed, three distinct types have beenrecognized:43 (1) commissural type, consist-ing of fusion of the commissures with littleinvolvement of cusps or chordae; (2) cuspaltype in which the leaflets are converted intostiff, rigid, leathery (later calcified) structures;and (3) chordal types in which the chordaeare fused, thickened, and shortened, therebyinterfering with the mobility of the leaflets. Inaddition to the pure forms, combinations ofthese types occur.The various anatomic forms of mitral

stenosis may affect atrioventricular filling insimilar manner. The degree of mitral valveobstruction is often fixed, due to commissuralfusion, and possibly, to the chordal abnormal-ities (fig. 1B). However, in the pure cuspalform (fig. IC) of mitral stenosis, the degree ofapparent valve narrowing, as evidenced clini-cally and hemodynamically, appears moresevere than that found anatomically. Thisphysiologic-anatomic dissociation is likely re-lated to stiffened and possibly calcified valvecusps; these cusps, while potentially mobile,may fail to open in response to a given leftatrial pressure, regardless of whether thecommissures are fused or not. It is therefore

883


SELZER, COHN

Figure 1Drawing of three m7iitral valves: (A) Normal initral valve in a closed position, viewed fromtire atrium. (B.) Stenotic m7iitral valve with corno7zisstral type of stenosis, showving maxinmmiir1iopening; commisstral fnision joins leaflets of normal thickness and mobility. (C.) CuspoltyJpe of mitral stenosis: onily rminimnal fiusion of comnmissucres is present, hut the stiff, fibro-calcific leaflets can.not open uniider- physiologic pressur.e.

possible to have the mitral valve "wide open"on surgical or pathologic inspection, and yetto have the valve remain severely stenoticunder in vivo conditions.

Progression of Mitral Stenosis:A New Concept

It is evident from the foregoing discussionithat clinical, hemodynamic, and pathologicobservation all point to the tact that mitralstenosis is often a continuously progressive,life-long disease. The clinical course showsthat most patients with mitral steniosis, whobecome disabled, do so relatively late in life,sometimes as late as in their fifties and sixties'hemodynamic studies revealed that suchpatients may show rapid progression in thedegree of mitral valve stenosis. Pathologicstudies indicate that pure commissural mitralstenosis with thin, mobile leaflets is foundmostly in young individuals, xvhile in olderpatients valve calcifications are almost alwayspresent. Onie should ask the question: Howdoes this course fit into the generally acceptedrheumatic origin of mitral steniosis? Very littleattention has been paid to the pathogenesis oflater stages of mitral stenosis. Initiated byacute rheumatic carditis, mitral stenosis isbelieved to be the consequence of healing ofthe rheumatic process. It is generally assumed

that active rheumatic process plays a role inthe progression of mitral valve changes, eithle-as repetitive rheumatic insults or as a "smol-dering" chronic rheumatic activity, botlh luisual-ly subelinical. This last view has receixved aboost by the frequent findings of Asehoffbodies in atrial muscle biopsies at the earlystage of the surgical era.

It should be pointed out that mans- kniown,facts about the natural history of iintralstenosis are not compatible with the -unitar-ian' rheumatic pathogenesis of mitral stenosis.If low-grade rhleumatic activity xere a prere-quisite to the scarring of mitral valve, then thepathologic process xw ould be most active 1in theearly years after the acute attack, aind thenixxwould gradually subside, for it is xellrecognized that both signs of rheumaticactivity and recurrences of rheumatic carditisare most constantly and frequently- foundduring the first two decades after the initialattack. Yet, our study has clearly demonstrat-ed that progression of mitral stenosis in thefourth and fifth decades of life often occurs ata rate that only an acceleration of the processcould explain. Otherxvise patients wxould nothave survived to that age. Fur thermiore,restenosis of the mitral valve, occurriing at any

Circultion, Volume XL. April 1972

884


MITRAL STENOSIS

age, would be difficult to fit into the activerheumatic concept.

Considering an alternate explanation forprogression of mitral stenosis beyond its earlystages, the most attractive hypothesis is toabandon the postulate of continuing rheuma-tic activity and to accept a theory that theinitial rheumatic valvulitis and its earlyscarring consequences produce changes uponthe valve that are capable later of perpetuat-ing themselves in a nonspecific manner. Thekey issue to this is the view proposed byEdwards44 in connection with calcific aorticstenosis, that a valve can be traumatized byabnormal flow patterns, and that such traumacan lead to thickening, fibrosis, and calcifica-tion of valve cusps. The clinical course ofcalcific aortic stenosis, now believed largelydue to congenitally abnormal cusps (mostlybicuspid), suggests that the narrowing of theaortic orifice is not a continuous, even process,but one likely to progress in an exponentialmanner. It may take 40 or 50 years oftraumatic damage to a bicuspid aortic valve toproduce barely discernible aortic stenosis; inthe next 5 to 10 years, the process mayaccelerate itself rapidly, producing a life-threatening, severe stenosis.The probability exists that an analogous

pathologic process accounts for the laterstages of mitral stenosis. The view that therheumatic process is the only mechanismproducing changes upon mitral valve cuspshas recently been contradicted by Pomer-ance,45 who found nodular thickening uponthe leaflets of elderly individuals, resemblingrheumatic changes, attributed to trauma ofvalve closure, the two leaflets hitting eachother. Similarly, mitral valve disease inLutembacher's syndrome (see above) may bedue to nonspecific trauma.One should thus visualize the rheumatic

process as one responsible for the earlychanges upon the leaflets leading eventually totheir fusion. Such fusion, producing thecommissural type of mitral stenosis, may causeonly minimal trauma or none at all, becausethe normal atrioventricular filling pathwaymay not be affected. In such cases, mild mitralCirculation, Volume XLV, April 1972

stenosis may persist for life. Slight alterationof flow pattern producing minimal trauma tothe valve may be analogous to a bicuspidaortic valve, i.e., produce progressive valvechanges over a period of many decades. Moresevere involvement of the commissures andthe involvement of the cusps and chordae bythe early rheumatic process would alter flowpatterns through the mitral orifice to a degreethat progressive valvular deformity wouldensue earlier in life. To produce clinically andhemodynamically significant mitral stenosis,further fusion of the commissures may takeplace, stiffening and calcification of the cuspswith or without fusion may develop, or a"funnel-type" deformity of the mitral valvemay take place, in which case major participa-tion of the chordae occurs, all of which maybe traumatic."

All observations regarding the natural his-tory of mitral stenosis can be explained by thenonspecific pathogenesis as proposed above,including the persistence of nonprogressivemild mitral stenosis for life, its late rapidprogression, its accelerated course in under-developed and depressed areas, the effect ofmitral valvotomy: permanent relief in some,delayed restenosis in others, immediate reste-nosis in still others.

Surgical and Postsurgical Stagesof Mitral Stenosis

The introduction of mitral valvotomy in1948 was a major breakthrough in thetreatment of mitral stenosis. The dramaticeffect of such an operation on seriously illpatients with mitral stenosis was immediatelyevident, but soon its limitations became ap-parent as well. The operation was clearly not acurative one, such as is the case in somecongenital cardiac lesions. It was limited topatients with pure or almost pure mitralstenosis, eliminating cases with all but trivialdegrees of regurgitation. Furthermore, itwas unsuccessful in some individuals, or evencapable of making the patients worse in a few(by the production of mitral regurgitation).Later developments included the performanceof valvotomy under direct vision by means of

88S


SELZER, COHN

open-heart surgery, and eventually, mitralvalve replacement.Even though more than 20 years has

elapsed from the earliest mitral valve opera-tions, the full impact of cardiac surgery uponthe natural history of mitral stenosis is as yetto be determined. It is obvious, however, thatin terms of risk, rate of complications, andlong-term fate of patients, mitral valvotomybelongs in an entirely different class frommitral valve replacement, and the two have tobe discussed separately.The evaluation of the results of mitral

valvotomy presents considerable difficulty.The overall beneficial results, in terms of largeseries of cases, can best be demonstrated bythe survival rate, which shows postvalvotomypatients, originally in functional classes IIIand IV, surviving a 10-year period at a higherrate than medically treated patients.17 How-ever, wide discrepancies exist between thevarious series in regard to the clinicalconditions of patients, as well as to the rate ofrestenosis of the mitral valve. It is highlyprobable that earlier studies, based on clinicalevaluation, overestimated the results ("excel-lent results" in 65-85% of cases).46 4 Laterreports agree that deterioration occurs fre-quently in patients with initial improvement.The incidence of restenosis of the mitral valve,as the basis for deterioration, was estimated aslow as 10%17' 46 and as high as 50%48 both in a10-year time period. Although the existenceand frequency of restenosis of the mitral valvehave been questioned from time to time49_since many such suspected cases were thoughtto represent merely unsuccessful valvoto-mies-hemodynamic studies indicate clearlythat restenosis does occur and may take placein those patients in whom mitral valvotomywidened the orifice only slightly, as well as inthose in whom the stenosis was nearly totallyrelieved.39Hemodynamic studies performed in only a

small fraction of patients included in thevarious clinical series revealed importantinformation. First, it is shown that mitralvalvotomy- may indeed produce a dramaticimprovement in circulatory dynamics, includ-

ing a regression of disappearance of reactivepulmonary hypertension.36 38 It was demon-strated, furthermore, that in many patientsvalvotomy produces a change from severe toonly slightly less severe mitral stenosis. Themajority of patients show a moderate reduc-tion of the mitral valvular obstruction, andonly a handful end up with an almostcomplete relief of mitral stenosis (entirelynormal dynamics under all conditions areprobably not attainable). Finally, a number ofpatients fail to show any improvement what-soever, even though the patient may have feltmuch better. Thus, subjective evaluation ofthe effect of the operation by the patientsneed not reflect the true anatomic success ofthe operation. For example, it has been shownthat abrupt deterioration in patients withmitral stenosis may develop in response to asmall increment in the severity of stenosis, asif a patient would "turn a corner" at a criticalpoint. Equally dramatic disappearance ofdisability can occur when the operationattained only a minimal success but managedto take the patient below his critical point ofstenosis.50How do pathologic findings correlate with

clinical and hemodynamic data? There seemsto be general agreement that mitral leafletswith good mobility offer the best target forconservative surgical approach. Thus, purecommissural type of mitral stenosis can besuccessfully relieved, in some cases perma-nently. It seems likely that invol-vement of thecusps and chordae in the stenotic processwould permit the attainment of only tempora-ry relief of mitral valve obstruction, withrestenosis becoming only a matter of time.Finally, severe fibrocalcific disease of the valvecusps may make relief of the obstructionimpossible, for here physiologic stenosis isproduced by inability of mitral leaflets -tomove under available filling pressure. Hence,it may become irrelevant to the degree ofresistance to flow whether the commissuresremain fused together or are cut all the Wayto the annulus. (Such cases illustrate the oc-casionally observed paradox of the surgeon

Circulation, Volume -XLV, April 1972

886


MITRAL STENOSIS

accomplishing a wide orifice, but hemody-namic studies showing no improvement.)

In response to the question whether mitralvalvotomy is still a worthwhile operation, theanswer appears to be an unequivocal yes. Thebeneficial effect of mitral valvotomy in proper-ly selected cases is well proven. The majorprerequisite is evidence of good mobility ofmitral cusps. On the other end of the spectrumof mitral stenosis are patients with heavycalcification of the valve cusps, in which casevalvotomy is ineffective. Regarding cases inbetween these two extremes, the choice ofoperation deserves most careful consideration.Equating the low risk (3-5% in the case ofclosed valvotomy and slightly higher in openvalvotomy), and the lack of late complica-tions, other than the possibility of restenosis,against the high risk in valve replacementwith a continuous risk of late complicationsand as-yet-unknown average life of the valve,a good case could be made in favor of theperformance of mitral valvotomy as a first tryin select cases. The alleged advantages ofopen valvotomy over closed valvotomy arebased entirely upon the surgical impressions.No "hard data" are available to compareresults of the two techniques. Hemodynamicfollow-up studies have amply demonstratedthe capability of closed valvotomy of achiev-ing spectacular improvement so that the valueof closed valvotomy in properly selected casesis established beyond doubt.The introduction of artificial mitral valve

constituted one of the breakthroughs ofcardiac surgery by providing the opportunityto save and rehabilitate many seriously illpatients with mitral stenosis nonresponsive tomore conservative surgical treatment. Never-theless, the effect of valve replacement uponthe natural history of mitral stenosis cannot beassessed as yet. What's more, the criticalstudent of the natural history of mitral stenosishas to ask the question whether the net effectof valve replacement may not be a shorteningrather than a lengthening of the life ofaverage patients with mitral stenosis. On thepositive side, hemodynamic studies after valvereplacement may show normal or near-normalCirculation, Volutne XLV, April 1972

values at rest, but abnormal responses toexercise are still the rule.5' The pulmonaryvascular resistance falls and the cardiac outputrises within hours after the operation.)} On thenegative side, artificial valves have not yetreached the point of resistanceless atrioven-tricular filling and are really equivalent to amild mitral stenosis. The risk of the operationis considerably higher than that of conserva-tive mitral surgery. Prosthetic mitral valvesintroduce the added risk of late complications;thromboembolic phenomena, periannularleaks, hemolysis, ball variance, and erosion ofthe cage through the wall of a small leftventricle (hence, preference of disc valves bysome surgeons). The possible need for anti-coagulant therapy introduces another riskfactor. To be sure, many technical advancesin the manufacture of prosthetic valves havebeen made and the use of homo- and hetero-grafts is being investigated. Yet, only time willultimately tell how many patients will haveuneventful 10 years of life after mitral valvereplacement and how often valve failure, ifpresent, will give enough warning time topermit reoperation to exchange valves.Perhaps the most revealing point about

present use of valve prosthesis is statistical,pertaining to the operative risk. In largecumulative series a fall in mortality of morerecent cases is emphasized, and figures of 4%mortality for mitral valve replacement areshown.53 4 Yet, a recent truly consecutiveseries in a cooperative study of six majorsurgical centers shows the overall hospitalmortality for mitral valve replacement of 22%.=.)It is pointed out that mortality figures of otheroperations in this cooperative study arecomparable with those of other centers.Therefore, surgical skill could not account forthe difference between a 4 and a 22% surgicalrisk of mitral valve replacement. A reasonablepossibility may be that the series with a 4%mortality contains a much higher number oflow-risk individuals with early forms of mitralstenosis, who would not be subjected to mitralvalve replacement in the other centers, or thatseriously ill patients with mitral stenosis arebeing turned down, or both! In centers

887


SELZER, COHN

participating in the cooperative study patientsare selected for surgery after hemodynamicevaluation. In many centers this is not thecase and only clinical evaluation is performed.While it is usually stated that valve replace-ment is performed on class III patients,functional classification is largely subjective,and the variability of the course of mitralstenosis makes a patient with rapid atrialfibrillation and class IV disability often returnto class I under adequate medical treatment.

Figures pertaining to late mortality aftervalve replacement are also difficult to inter-pret. Such figures are estimated at from 7 to20%. Assuming a 15% late mortality and a 22%immediate risk, the 5-year survival of abouttwo thirds is similar to the average 70%survival in the series of Rowe et al.15 dealingwith medical therapy. The argument thatearly operations lower the risk of valvereplacement56 is weakened by the fact thatlate deaths, complications, and the ultimatefate of the patients have not been demonstrat-ed to differ in those operated at early stagesfrom those operated in later stages of mitralstenosis, so that medical treatment may carryan overall lower risk.

General DiscussionThe main thesis presented here is based on

indirect evidence, but is consistent with allknown facts about the course and the naturalhistory of mitral stenosis. The view isproposed that mitral stenosis is initiated by arheumatic insult to the mitral valve, whichaccounts for the early changes. The laterprogression of the anatomic changes of themitral valve is a nonspecific process resultingfrom trauma to the valve structures caused byaltered flow patterns, analogous to the nowuniversally accepted mechanism of calcificaortic stenosis developing upon a bicuspidaortic valve or one showing other congenitaldeviation from the norm. It is believed thatthe progression of mitral stenosis does notoccur in a uniform fashion, but goes through astage of stability, which at some point turnsinto progressive stage, often at an accelerated

rate. Severe rheumatic insults, such as thoseseen in underdeveloped parts of the world,may produce such accelerated progressionearly in life. In our area this stage is likely tooccur in the fourth, fifth, or sixth decade oflife. Patients may also remain unchanged for alifetime with mild or moderate mitral stenosis,if more favorable anatomic conditions exist.The application of this thesis to the

management of patients with mitral stenosisjustifies a conservative approach and caution,especially when committing the patient tosuch an irrevocable form of surgical treat-ment as valve replacement. Many misconcep-tions regarding mitral stenosis are prevalentamong physicians and the public alike. Thepatient's primary physicians often believe thatmitral stenosis is a "surgical disease" and thathe should be channeled directly to a cardiacsurgeon. Patients may be under the impressionthat they suffer from a condition for whichthere is a "surgical cure," and begin to pressfor an operation whenever the slightestinconvenience related to mitral stenosis oc-curs. If we are to influence the natural historyof mitral stenosis in a most favorable way, it isnecessary to follow the team approach, withexperienced clinicians, physiologists, and sur-geons critically reviewing each case individ-ually. Medical treatment and surgical treat-ment are two parallel modes of managementof mitral stenosis, the choice of which needs tobe carefully defined. Surgical candidates musthave reasonable hemodynamic evidence ofsignificant valve obstruction. Clinical symp-tomatology should be clearly related to mitralstenosis per se, rather than to some of itssecondary complications, preferably should beprogressive (the value of serial hemodynamicevaluations is self-evident), and nonrespon-sive to a good medical regimen.Further studies are needed to provide more

detailed information regarding the pathologicstage of mitral stenosis and its hemodynamicprogression, as well as angiocardiographicidentification of its varieties. Such studiesmight help provide rational and scientificallybased guidelines for the timing and type of


888


MITRAL STENOSIS

surgical treatment that would most favorablyaffect the course of the various forms of mitralstenosis.

AcknowledgmentThe authors are indebted to Dr. Jesse E. Edwards

for reviewing sections of this article dealing withpathology of the mitral valve.

References1. POYNTON FJ, PAYNE A: The aetiology of rheu-

matic fever. Lancet 2: 861, 19002. COBURN AF: The Factor of Infection in the Rheu-

matic State. Baltimore, Williams & WilkinsCo., 1931

3. RICK AR, GREGORY JE: Experimental evidencethat lesions with the basic characteristics ofrheumatic carditis can result from anaphylactichypersensitivity. Bull Johns Hopkins Hosp 73:239, 1943

4. MURPHY GE, SWIFT HF: Induction of cardiaclesions, closely resembling those of rheumaticfever in rabbits following repeated skininfections with Group A streptococci. J ExperMed 89: 687, 1949

5. GOLDSTEIN I, HALPERN B, ROBERT L: Immuno-logical relationship between streptococcuspolysaccharide and the structural glycoproteinsof heart valve. Nature 213: 44, 1967

6. ALBA B, EPSTEIN JA, FEINSTEIN AR, GAVRINJB, SARAN J, KLEINBERG E, SIMPSON R,TARANTA A, TURSKY E, WOOLF HF: Rheumaticfever in children and adolescents: A long-termepidemiological study of subsequent prophylax-is, streptococcal infections and clinical se-quelae. Ann Intern Med 60 (suppl 5): 3,1964

7. BURCH GE, COLCOLOUGH HL: Viral valvulitis.Amer Heart J 78: 119, 1969

8. FERENCZ C, JOHNSON AL, WIGLESWORTH FW:Congenital mitral stenosis. Circulation 9: 161,1954

9. STEINBRUNN W, COHN KE, SELZER A: Atrialseptal defect associated with mitral stenosis:The Lutembacher syndrome revisited. Amer JMed 48: 295, 1970

10. EDWAIRDS JE: An Atlas of Acquired Diseases ofthe Heart and Great Vessels. Philadelphia, W.B. Saunders Co., 1961, p 56

11. OKADA R, GLAGOV S, LEV M: Relations of shuntflow and right ventricular pressure to heartvalve structure in atrial septal defect. AmerHeart J 78: 781, 1969

12. CooMBs CF: Rheumatic Heart Disease. NewYork, William Wood and Co., 1924

13. BLAND EF, JONES TD: Rheumatic fever andrheumatic heart disease. A twenty-year reporton 1000 patients followed since childhood.Circulation 24: 836, 1961


14. WOOD P: An appreciation of mitral stenosis. BritMed J 1: 1051, 1113, 1954

15. ROWE JC, BLAND EF, SPRAGUE HB, WHITE PD:Course of mitral stenosis without surgery: Tenand twenty perspectives. Ann Intern Med 52:741, 1960

16. OLESEN KH: The natural history of 271 patientswith mitral stenosis under medical treatment.Brit Heart J 24: 349, 1962

17. ELLIs LB, HARKEN DE: Closed valvuloplasty formitral stenosis: A twelve year follow-up studyof 1571 patients. New Eng J Med 270: 643,1964

18. ANGELINO PF, LEVI V, BRUSCA A, ACTIS-DATOA: Mitral commissurotomy in younger agegroup. Amer Heart J 51: 916, 1956

19. BORMAN JB, STERN S, SHAPIRA T, MILVIDSKY H,BRAUN K: Mitral valvotomy in children. AmerHeart J 61: 763, 1961

20. CHERIAN G, VYTILINGAM KI, SUKUMAR IP,GOPINATH N: Mitral valvotomy in youngpatients. Brit Heart J 26: 157, 1964

21. AL-BAHRANI IR, THAMER MA, AL-OMERI MM,AL-NAAMAN YD: Rheumatic heart disease inthe young in Iraq. Brit Heart J 28: 824,1966

22. ROY SB, GOPINATH N: Mitral stenosis. Circula-tion 38 (suppl V): V-68, 1968

23. MANTEUFFEL-SZOEGE L, Nowicn J, WASNIEWSKAM, SITKOWSKI W, TURSKI C: Mitral commis-surotomy: Results of 1700 cases. J CardiovascSurg 11: 350, 1970

24. GLOVER RP: Mitral stenosis in a young girl. AnmerJ Cardiol 4: 132, 1959

25. LOGAN A, TURNER R: Mitral stenosis: Diagnosisand treatment. Lancet 1: 443, 1953

26. BAILEY GWH, BRANIFF BA, HANCOCK EW,COHN KE: Relation of left atrial pathology toatrial fibrillation in mitral valvular disease. AnnIntern Med 69: 13, 1968

27. SELZER A: Effects of atrial fibrillation upon thecirculation in patients with mitral stenosis.Amer Heart J 59: 518, 1960

28. ELLIS LB, HARKEN DE: Arterial embolization inrelation to mitral valvuloplasty. Amer Heart J62: 611, 1961

29. CASELLA L, ABELMANN WH, ELLIS LB: Patientswith mitral stenosis and systemic emboli. ArchIntem Med (Chicago) 114: 773, 1964

30. SZEKELY P: Systemic embolism and anticoagulantprophylaxis in rheumatic heart disease. BritMed J 1: 1209, 1969

31. FLEMMING HA, WOOD P: The myocardial factorin mitral valve disease. Brit Heart J 21: 117,1959

32. HARVEY RM, FERRER I, SAMET P, BADER RA,BADER ME, COURNAND A, RICHARDS DW:

889


SELZER, COHN

Mechanical and myocardial factors in rheumat-ic heart disease with mitral stenosis. Circula-tion 11: 531, 1955

33. WOOD P: Diseases of the Heart and Circulation,ed 2. Philadelphia, J. B. Lippincott Co., 1956

34. RAMSEY HW, DE LA TORRE A, BARTLEY TD,LINHART JW: Intractable hemoptysis in mitralstenosis treated by emergency mitral commis-surotomy. Ann Intern Med 67: 588, 1967

35. GORLIN R, GORLIN SC: Hydraulic formula forcalculation of the area of the stenotic mitralvalve, other cardiac valves, and centralcirculatory shunts. Amer Heart J 41: 1,1951

36. SELZER A, MALMBORG RO: Some factors influenc-ing changes in pulmonary vascular resistance inmitral valvular disease. Amer J Med 32: 532,1962

37. ARMSTRONG WT, SELZER A: Mitral stenosis withsevere pulmonary hypertension. To be pub-lished

38. REEVE R, SELZER A, POPPER RW, LEEDS RF,GERBODE F: Reversibility of pulmonary hyper-tension following cardiac surgery. Circulation33 (suppl I): I-107, 1966

39. DUBIN A, COHN KE, MARCH HW, SELZER A:Longitudinal hemodynamic and clinical studyof mitral stenosis. Circulation 44: 381 1971

40. HUDSON REB: Cardiovascular Pathology. Balti-more, Williams & Wilkins Co., 1965

41. MAGAREY FR: Pathogenesis of mitral stenosis.Brit Med J 1: 856, 1951

42. BROCK RC: The surgical and pathologicalanatomy of the mitral valve. Brit Heart J 14:489, 1952

43. RUSTED IE, SCHEIFLEY CH, EDWARDS JE:Studies of the mitral valve: II. Certainanatomic features of the mitral valve andassociated structures in mitral stenosis. Circula-tion 14: 398, 1956

44. EDWARDS JE: The congenital bicuspid aorticvalve. Circulation 23: 485, 1961

45. POMERANCE A: Aging changes in human heartvalves. Brit Heart J 29: 222, 1967

46. BAKER C, HANCOCK EW: Deterioration aftermitral valvotomy. Brit Heart J 22: 281,1960

47. ELLIS LB, HARKEN DE, BLACK H: Clinical studyof 1000 consecutive cases of mitral stenosis twoto nine years after mitral valvuloplasty.Circulation 19: 803, 1959

48. TURNER RDW: Surgical treatment of mitralstenosis. Postgrad Med 46: 116 (Sept),1969

49. HIGGS LM, GLANCY DL, O'BREN KP, EPSTEINSE, MORROW AC: Mitral restenosis: Anuncommon cause of recurrent symptoms fol-lowing mitral commissurotomy. Amer j Cardiol26: 34, 1970

50. SELZER A, POPPER RW, GERBODE F, KERTH WJ:Effect of surgical relief of cardiac overloadupon intractable heart failure. Amer J Med Sci251: 183, 1966

51. MoRRow AG, OLDHAM HN, ELKINS RC,BRAUNWALD E: Prosthetic replacement of themitral valve: Preoperative and postoperativeclinical and hemodynamic assessments in 100patients. Circulation 35: 962, 1967

52. DALEN JE, MATLOFF JM, EVANS GL, HOPPINFG JR, BHARWAJ P, HARKEN DE, DEXTER L:Early reduction of pulmonary vascular resis-tance after mitral valve replacement. New EngJ Med 277: 387, 1967

53. HODAM R, STARR A, RAIBLE D, GRIswOLD H:Totally cloth-covered prostheses: A review oftwo years' experience. Circulation 41 (supplII): II-33, 1970

54. BEALL AC, BLOODWELL RD, BRICKER DL, OKIESJE, COOLEY DA, DEBAKEY ME: Prostheticreplacement of cardiac valves: Five and onehalf years' experience. Amer J Cardiol 23:150, 1969

53. KITTLE CF, DYE WS, GERBODE F, GLENN WWL,JULIAN OC, MoRRow AG, SABISTON DC,WEINBERG M: Factors influencing risk incardiac surgical patients: Cooperative study.Circulation 40 (suppl I): I-169, 1969

56. BRYANT LR, TRINKLE JK, SPENCER FC,DANIELSON GK, SHABETAI R, REEVES JT:Cardiac valve replacement. Results in patientswith advanced disability. JAMA 216: 996,1971


890


ARTHUR SELZER and KEITH E. COHNNatural History of Mitral Stenosis: A Review

Print ISSN: 0009-7322. Online ISSN: 1524-4539 Copyright 1972 American Heart Association, Inc. All rights reserved.

75231is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TXCirculation

doi: 10.1161/01.CIR.45.4.8781972;45:878-890Circulation.

http://circ.ahajournals.org/content/45/4/878.citationlocated on the World Wide Web at:

The online version of this article, along with updated information and services, is

http://circ.ahajournals.org//subscriptions/is online at: Circulation Information about subscribing to Subscriptions:

http://www.lww.com/reprints Information about reprints can be found online at: Reprints:

document. Permissions and Rights Question and Answer of the Web page under Services. Further information about this process is available in thewhich permission is being requested is located, click Request Permissions in the middle columnClearance Center, not the Editorial Office. Once the online version of the published article for

can be obtained via RightsLink, a service of the CopyrightCirculationoriginally published in Requests for permissions to reproduce figures, tables, or portions of articlesPermissions:


circulation 1972 selzer 878 90

Documents