|
|
| |||||||||||||||||||||||||||||||||||||||||
Background Valvular heart disease is the most important cardiac manifestation of systemic lupus erythematosus. We performed a study to determine the relation of valvular disease to other clinical features of lupus, whether or not the valve disease progresses, and the associated morbidity and mortality.
Methods We performed transesophageal echocardiography and rheumatologic evaluations in 69 patients with systemic lupus erythematosus. The echocardiographic findings were compared with those in 56 healthy volunteers. Fifty-eight patients (84 percent) had second evaluations a mean (±SD) period of 29±13 months later. The patients and controls were followed for 57 months.
Results Valvular abnormalities were common on the initial and the follow-up echocardiograms (in 61 and 53 percent of the patients, respectively). Valvular thickening was the predominant finding initially and on follow-up (in 51 and 52 percent of the patients, respectively), followed by vegetations (in 43 and 34 percent), valvular regurgitation (in 25 and 28 percent), and stenosis (in 4 and 3 percent). Valvular abnormalities frequently resolved, appeared for the first time, or persisted but changed in appearance or size between the two studies. Mild or moderate valvular regurgitation did not progress to become severe, and new stenoses did not develop. Neither the presence of valvular disease nor changes in the echocardiographic findings were temporally related to the duration, activity, or severity of lupus or to its treatment. The combined incidence of stroke, peripheral embolism, heart failure, infective endocarditis, and the need for valve replacement was 22 percent in the patients with valvular disease, but only 8 percent in those without it. A total of seven patients died during follow-up, in most cases as a result of valvular disease. Valvular abnormalities and complications were uncommon in the controls (occurring in 9 and 2 percent, respectively).
Conclusions Valvular heart disease is common in patients with systemic lupus erythematosus, frequently changes over time, appears to be temporally unrelated to other clinical features of lupus, and is associated with substantial morbidity and mortality.
Methods
Study Population
From 1989 to 1996, we performed transesophageal echocardiography and rheumatologic evaluations in 69 patients with systemic lupus erythematosus. The initial echocardiographic findings in 54 of these patients have been reported elsewhere.4 The 69 patients represented 86 percent of the outpatients 60 years of age or younger who are followed by rheumatologists at our affiliated hospitals. Patients over 60 years old were excluded because of the high prevalence of degenerative valvular disease in this age group.8 Patients with a history of rheumatic fever or intravenous drug abuse were also excluded.
There were 61 women and 8 men in the study, with a mean age of 38 years (range, 16 to 57). A total of 58 patients (84 percent) underwent a repeated evaluation a mean (±SD) of 29±13 months (range, 6 to 54) after the initial evaluation. Forty-two of the 58 were patients with stable lupus who underwent both evaluations as outpatients, with the second evaluation performed at least 12 months after the first. Sixteen patients had at least one hospital admission during the study period due to flares of the lupus or complications potentially related to valvular disease. The second evaluation was performed during hospitalization in 7 of the 16 patients, and within one month after discharge in 7. Initial echocardiographic findings were compared with those in 56 age-matched healthy volunteers (31 women and 25 men) with a mean age of 35 years (range, 17 to 54). The study was approved by the Human Research Committee at our institution, and all subjects provided informed consent.
Echocardiographic Studies
All subjects underwent initial and follow-up transesophageal echocardiography with a color Doppler imaging system (Hewlett–Packard Sonos 500, 1500, or 2500). All echocardiograms were coded according to the patient's or control's Social Security number, and the studies in patients with lupus were interpreted in random order along with those in healthy volunteers. The echocardiograms were interpreted by an observer who was unaware of the clinical data and the timing of the studies. In addition to standard echocardiographic views,9 images of the heart valves were obtained at high magnification in multiple planes.
Electronic calipers were used to determine the thickness of the mitral, tricuspid, and aortic valves on transesophageal echocardiography in the patients and controls. The basal, middle, and tip portions of the atrioventricular-valve leaflets were measured with M-mode echocardiography (for the mitral valve) and two-dimensional echocardiography (for the tricuspid valve) in the four-chamber view. The midportion of the right and noncoronary cusps of the aortic valve was measured by M-mode echocardiography from the basilar short-axis and longitudinal views. The left coronary cusp was measured by two-dimensional echocardiography from the basilar short-axis view. In the normal subjects, the thickness of the mitral and tricuspid valves was 0.7 to 3 mm, and the aortic-valve thickness was 0.7 to 2 mm. Abnormal valvular thickening was therefore considered to be present when a thickness of more than 3 mm (for the mitral and tricuspid valves) or more than 2 mm (for the aortic valve) was observed in at least two leaflets or in one leaflet with an associated vegetation, at least mild regurgitation, or both. The pulmonic valve was not measured, because thickening of this valve was rare.
A valvular vegetation was defined as an abnormal localized echodensity with well-defined borders either part of or adjacent to valve leaflets, the subvalvular apparatus, or the great vessels. The size of the vegetation was determined by planimetry and expressed in square centimeters.
Mitral- or tricuspid-valve regurgitation was graded with the use of the color Doppler jet-area method.10 Aortic regurgitation was graded according to the ratio of the width of the color jet to the diameter of the outflow tract.11 This method was derived from transthoracic studies and applied to transesophageal echocardiography. Pulmonic regurgitation was assessed visually. Regurgitation was considered to be present if it was more than mild or if it was mild and associated with valvular thickening or a vegetation of the respective valve. Minimal or mild regurgitation of a structurally normal valve was not documented.
The severity of valvular stenosis was assessed by calculating the valvular area, with the use of the continuity equation.12
Clinical Evaluation
All patients underwent a complete clinical evaluation at the time of each echocardiographic study. The activity of the lupus was assessed with the Lupus Activity Criteria Count.13 The disease was considered to be active if at least two of seven clinical and laboratory criteria were met. The severity of the disease was graded with the Lupus Severity of Disease Index14 and scored on a scale from 1 to 13.
A laboratory-activity score of 1 to 10 was derived from 10 laboratory abnormalities: a positive test for antinuclear antibodies (dilution, >1:40), a positive test for DNA antibodies (dilution, >1:10), anemia (<12 g of hemoglobin per deciliter), leukopenia (<4000 white cells per cubic millimeter), thrombocytopenia (<150,000 platelets per cubic millimeter), renal failure (>1.4 mg of creatinine per deciliter [>120 µmol per liter]), hematuria (>5 red cells per high-power field), a low C3 value (<85 mg per deciliter), a low C4 value (<12 mg per deciliter), and a high erythrocyte sedimentation rate (>20 mm per hour).
The average dose of each drug taken during the three months preceding each echocardiographic study was assessed by chart reviews or personal or telephone interviews with the patients.
Morbidity and Mortality
Patients and controls were followed for 57±12 and 57±15 months, respectively. The incidences of stroke or peripheral embolism, heart failure, infective endocarditis, need for valve replacement, and death were determined by interviews with the patients and chart reviews.
Statistical Analysis
The frequency of and degree of change in valvular abnormalities in the patients and controls were compared with Fisher's exact test.15 Paired comparisons for binary variables were performed with McNemar's test.15 Student's and paired t-tests were used for the comparison of continuous variables.16 To assess interrater agreement in the detection of valvular abnormalities, a second independent observer interpreted 45 randomly selected echocardiograms from patients and controls. Interrater agreement was determined by the percentage of agreement between the two sets of readings, as well as by Cohen's kappa test.17 A two-tailed P value of less than 0.05 was considered to indicate statistical significance.
Results
Valvular Abnormalities
In the patients with lupus, valvular abnormalities — predominantly of the mitral and aortic valves — were common on the initial echocardiogram (in 61 percent of the patients) and on the follow-up echocardiogram (in 53 percent). Abnormalities were present in only 9 percent of the controls (P<0.001 for the comparison with the patients, for each abnormality except valvular stenosis) (Table 1).
|
Valvular vegetations were present in 30 of the 69 patients (43 percent) initially and in 20 of 58 (34 percent) at follow-up. The vegetations were located on the basal, middle, or tip portions of the leaflets and were located predominantly on the atrial side of the mitral valve or on the vessel side of the aortic valve. Vegetations were of variable size (range, 0.06 to 1.25 cm2) and shape, had irregular borders, and frequently had heterogeneous echogenicity (Figure 1A, Figure 1B, and Figure 1C). Two thirds of the patients with valvular vegetations had associated valvular thickening.
|
Over half of the 45 echocardiograms evaluated by a second reader showed evidence of valvular disease. Overall, there was 96 percent agreement (kappa = 0.91) between the two sets of readings. Agreement in detecting valvular vegetations was 98 percent for the aortic valve and 96 percent for the mitral valve. Similar results were obtained for valvular regurgitation (93 and 98 percent, respectively) and thickening (87 and 93 percent, respectively).
Nonvalvular Cardiac Diseases
Fifteen of the 69 patients (22 percent) had nonvalvular cardiac abnormalities. Four patients had segmental left ventricular dysfunction due to coronary artery disease, and five had global hypokinesis of unknown cause. Five patients had pericardial effusions, and one of the five had fatal cardiac tamponade. One patient had severe cor pulmonale due to recurrent pulmonary embolism.
Relation of Valvular Disease to Clinical Features of Lupus
The presence of valvular disease was not temporally related to the clinical features of lupus. The patient's age, the duration of the lupus, clinical or laboratory evidence of its activity and severity, and the duration or type of therapy did not differ according to the presence or absence of valvular abnormalities (Table 2).
|
Valvular abnormalities, especially vegetations, frequently resolved, appeared for the first time, or persisted but with changes in appearance or size during the period between the initial and follow-up echocardiograms (Table 3). In 14 patients (24 percent), 25 vegetations resolved, and in 7 patients (12 percent), 8 new vegetations were present at follow-up (Figure 1A, Figure 1B, and Figure 1C). Thirteen patients (22 percent) had 19 vegetations initially but 27 at follow-up. In 8 of these 13 patients, some vegetations resolved, some appeared for the first time at follow-up, and some persisted but with changes in size or appearance between the initial and follow-up echocardiographic studies.
|
The clinical manifestations of lupus did not change between the two evaluations and did not differ according to whether valvular abnormalities resolved, appeared for the first time, or persisted at follow-up (Table 2).
Morbidity and Mortality
During almost five years of follow-up, 29 complications developed in 12 of 58 patients (21 percent), including 7 deaths (12 percent) (Table 4). Of the 10 patients with valvular disease and complications, 5 had 6 ischemic strokes and 1 peripheral embolism. Of these five patients (three of whom had valvular vegetations and two of whom had valvular thickening with regurgitation), three had multiple cerebral infarcts on magnetic resonance imaging at the time of the complications (Figure 1A, Figure 1B, and Figure 1C). A sixth patient had a fatal hemorrhagic stroke while receiving anticoagulant therapy for a prosthetic valve. Five of the six patients with heart failure had severe valvular regurgitation, complicated by coronary artery disease in two and by infective endocarditis in two. Two of the patients underwent successful valve replacement, two declined surgery (both died from refractory heart failure), and one died from sepsis. In the sixth patient, right heart failure developed as a result of recurrent pulmonary embolism. Three patients had infective endocarditis of the mitral valve. In one of these patients, pathological studies confirmed the presence of both infective and Libman–Sacks vegetations on the mitral valve and Libman–Sacks vegetations on the tricuspid valve. This patient subsequently died of sepsis.
|
Discussion
In our study, neither the presence of valvular disease nor changes in valvular disease over time were temporally associated with the duration, activity or severity, or treatment of lupus. There are several possible explanations for the absence of any associations among these variables. The clinical and immunologic markers of activity in other organ systems may be present at a different time or have a different duration from those of valvulitis. Another possibility is that antiphospholipid antibodies, which are frequently present in patients with lupus independently of the activity or severity of the disease, may cause valvular injury.2,3,18 Furthermore, the many scoring systems based on clinical or laboratory variables for the assessment of the activity and severity of lupus may be limited in their accuracy.13,14,19,20,21,22
An important finding in this study is that over time, valvular abnormalities frequently resolve or change, and new abnormalities often develop. Valvulitis, like inflammation in other organ systems, is probably intermittent. Several pathological studies of patients with lupus have shown active and healed valvulitis, as well as active Libman–Sacks vegetations with acute thrombus, healed vegetations with or without hyalinized thrombus, or both active and healed vegetations, in the same or different valves.23,24,25 Thrombotic vegetations due to a hypercoagulable state have also been demonstrated in patients with lupus.26,27,28,29 These vegetations cannot be clearly differentiated from Libman–Sacks vegetations on echocardiography.30,31 In addition, valvular vegetations can embolize, which may result in a change in their appearance or resolution.7,28,29,32
In our study, most of the complications were related to valvular disease. All five patients with ischemic strokes or peripheral embolism had valvular disease. Cerebral vasculitis was suspected in one patient and confirmed in another, but both patients also had valvular vegetations. The patient with pathological confirmation of cerebral vasculitis had multiple small cerebral infarcts and both infective and Libman–Sacks vegetations on the mitral valve. Cerebral vasculitis and cardioembolism may therefore have coexisted in these two patients. Cardioembolism was the more likely diagnosis in the other three patients with stable disease, because cerebritis and vasculitis usually occur in patients with highly active disease. Other echocardiographic and pathological studies have reported a 10 to 20 percent incidence of ischemic stroke in patients with lupus,7,29,32 with valvular vegetations, valvulitis, or left-heart thrombus identified as the cardioembolic substrate in 70 to 90 percent of the patients. None of the patients in these studies had cerebral vasculitis,29,32 and none of the patients treated with anticoagulants had recurrent events.7
Infective endocarditis was an important cause of morbidity and mortality in three of our patients. It was the initial presentation in two patients and was complicated by severe valvular dysfunction, heart failure, and death from sepsis in two. The three patients with endocarditis had had disease flares requiring high doses of corticosteroids and cytotoxic agents shortly before their infections developed. Infection is the most common cause of death in patients with lupus, and the risk is highest among those with active disease treated with high doses of corticosteroids or cytotoxic agents.33,34 Antibody-mediated granulocyte dysfunction, leukopenia, hypocomplementemia, abnormal immunoglobulin synthesis, and functional asplenia are possible pathogenic factors.33,34,35 Infection can also mimic, accompany, or trigger a flare of the disease.
The mortality rate at five years and the causes of death (cardiac, cerebrovascular, and infectious) in our study are similar to those in other studies.1,2,3,32,36,37,38 In our study, however, heart disease was responsible for six of the seven deaths: severe valvular and coronary artery disease in two patients, infective endocarditis in two, cardiac tamponade in one, and intracerebral bleeding during anticoagulant therapy for a prosthetic valve in one. We cannot rule out the possibility that a selection bias in the study population contributed to this association.
Our results differ from those of studies using transthoracic echocardiography.1,2,3,32,36,37 In these studies, the prevalence of valvular disease was less than 40 percent, which may be related to the use of a less sensitive echocardiographic technique. Valvular abnormalities were also reported to be persistent or progressive in these studies. The inclusion of patients older than 60 years of age or a higher proportion of inpatients may explain this finding. As in our study, four of these studies found no association between valvular disease and indicators of the activity of lupus.1,32,36,37
The clinical course of lupus frequently fluctuates between periods of remission and flares, and valvular disease may parallel these clinical changes. The treatment of lupus is also generally initiated or modified according to its clinical features. Our study yielded an unexpected finding: neither the usual clinical and immunologic markers of the activity of lupus nor its treatment appeared to be temporally related to the presence of or changes in valvular disease. Therefore, in a patient with stable lupus, newly diagnosed valvular disease may not reflect an increase in the activity or severity of the lupus, and there may be no need to modify antiinflammatory therapy. In addition, the demonstration of valvular disease at one point in time is not predictive of permanent or progressive valvular dysfunction. Given the poor prognosis associated with infective endocarditis in patients with lupus, however, this complication should be looked for in febrile patients with active disease and clinically suspected or known valvular disease.
Since most patients with lupus and valvular disease have no cardiac symptoms, a careful cardiovascular examination should be the primary method of screening for valvular regurgitation or stenosis in patients who may require echocardiography or antibiotic prophylaxis for infective endocarditis. In addition, since strokes in patients with lupus are frequent and valvular thickening and vegetations are prevalent and can serve as substrates for cardioembolism, prophylactic antiplatelet therapy may be appropriate for such patients, with the recognition that there is no evidence supporting the efficacy of antiplatelet therapy in this setting.
Long-term anticoagulation should be considered independently of the echocardiographic findings in patients with focal neurologic deficits or peripheral embolism and no features of systemic vasculitis. In contrast, if vasculitis or cerebritis is suspected in a patient with active disease, transesophageal echocardiography can be used to detect a cardioembolic substrate, and the findings can be used as the basis for treatment with anticoagulants instead of or in addition to high-dose corticosteroids and cytotoxic agents. Finally, in patients with lupus and recurrent systemic embolism, the surgical excision of uninfected valvular vegetations may not prevent their recurrence.
We are indebted to Rachel Chaney, M.D., for help collecting clinical data; to Clifford Qualls, Ph.D., for statistical consultation; to Kristine Councilman, Aggie Shaeffer, and Frank Gurule for echocardiographic technical assistance; and to the Department of Rheumatology for referring patients.
Source Information
From the University of New Mexico Health Sciences Center and Veterans Affairs Medical Center, Albuquerque.
Address reprint requests to Dr. Roldan at the Cardiology Division, 5B-111, Veterans Affairs Medical Center, 2100 Ridgecrest Dr. SE, Albuquerque, NM 87108.
References
| |||||||||||||||||||||||||||||||||||||||||
Related Letters:
Valvular Heart Disease and Systemic Lupus Erythematosus
Piette J.-C., Amoura Z., Papo T., McCarthy G. M., McCarty D. J., Roldan C. A., Crawford M. H.
Extract |
Full Text
N Engl J Med 1997;
336:1324-1325, May 1, 1997.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2010 Massachusetts Medical Society. All rights reserved. |
Previous
