Functional tricuspid regurgitation (TR) is relatively common in patients with mitral or aortic valvular heart disease. Significant TR does not necessarily regress after surgical correction of mitral valvular lesions,1^–5 and so many physicians and surgeons advocate initial annuloplasty with or without a ring for moderate-to-severe TR at the time of corrective surgery of the left-sided valvular disease. In addition, significant TR can develop late after successful left-sided valve surgery, and this has become a major issue since the early reports of poor surgical outcomes.6^–10

Although long-term follow-up studies have sought to evaluate the incidence and risk factors associated with development of late TR, such studies have included organic TR patients, or only patients with rheumatic valvular disease or mitral valvular lesions.9 11 12 Thus, the impact of different aetiologies and/or diverse surgical procedures has not been adequately investigated, and, moreover, the real clinical outcomes of late TR have not been assessed. The present study investigated the incidence, risk factors and clinical impact of late TR after open-heart surgery for left-sided valvular heart disease.

METHODS

Subjects

Clinical data from patients who underwent mitral and/or aortic valve surgery between 1995 and 2000 at the Asan Medical Center were reviewed. A total of 870 patients underwent successful open-heart surgery for left-sided valvular lesions. Of those, the study excluded 192 patients who had undergone concomitant tricuspid repair surgery due to organic tricuspid valvular disease or a TR grade ⩾3/4 according to preoperative echocardiography. A further 40 patients were excluded due to clinical or echocardiographic follow-ups of less than 12 months. Thus, the study population consisted of 638 patients (282 women, 44%) aged 52 (14) (mean (SD)) years, and no patient showed any evidence of rheumatic involvement of the tricuspid valve on preoperative echocardiography.

Echocardiography

Prior to surgery, patients underwent comprehensive echocardiographic evaluation using a Hewlett-Packcard Sonos 2500 or 5500 imaging system equipped with a 2.5 MHz transducer. The maximal TR jet area visualised using colour Doppler flow mapping was used for TR quantification, with a TR jet-to-right atrial area ratio of <10% = grade 1+, 10–20% = 2+, 20–40% = 3+ and >40% = 4+.13 Systolic reversal of hepatic venous flow was also used as a criterion for 4+ TR. The peak systolic TR jet velocity measured by CW Doppler was used to calculate the TR peak pressure gradient using the simplified Bernouii equation (pressure gradient = 4(velocity)2) to estimate the peak systolic pulmonary artery pressure. Evidence of rheumatic valvular heart disease was based on careful evaluation of valvular morphology. A diagnosis of rheumatic valvular heart disease was based on the presence of commissural fusion and multivalvular involvement with thickening, in addition to motion limitation of the posterior mitral leaflet. Such diagnoses were confirmed during surgery. In 460 patients (72%), tricuspid annulus diameter was retrospectively measured in the low parasternal or apical four-chamber view in late diastole as the distance between the points of reflection of the septal and mural endocardium on the anterior and septal tricuspid leaflets.

Follow-up

Clinical follow-up was performed at 1, 2 and 6 months after surgery and annually thereafter. Clinical data were collected by a single cardiologist (MJK) through either a chart review or telephone interview. Clinical events included cardiac death, repeated open-heart surgery and congestive heart failure requiring hospital admission. Echocardiography was performed within 1 week of surgery (before discharge) and 6 months after surgery, after which an annual examination was recommended. Development of late significant TR was defined as a TR increase by more than one grade and final TR grade ⩾3+ on follow-up echocardiography; mild increase in TR from 2+ at baseline to 3+ at follow-up was not included as development of late significant TR. The mean clinical and echocardiographic follow-up durations were 101 (24) (12–146) and 64 (30) (12–141) months, respectively, and these were completed in December 2006. The clinical and echocardiographic follow-up rates were 100% and 96%, respectively.

Statistical analysis

All data are expressed as mean (SD). Differences between continuous variables for patients who did and did not develop late significant TR were analysed using the Student unpaired t test. Differences between categorical variables were compared using the χ² or Fisher exact test. Cox proportional hazard model was used for multivariate analysis of clinical variables that showed significant between-group differences (ie, age, gender, rhythm, rheumatic aetiology, surgical procedure, left atrial size, initial TR grade, peak pressure gradient of TR at follow-up and prosthesis malfunction). All statistical analyses were performed using SPSS version 12.0 for Windows (SPSS, Chicago), and the null hypothesis was rejected for p values <0.05.

RESULTS

Incidence of late significant TR

Of the total 638 patients, 544 underwent single valve surgery, comprising 323 who underwent surgery for mitral valvular heart disease and 221 for aortic valvular heart disease. The underlying disease was mitral regurgitation in 198 patients and mitral stenosis in 102. For aortic-valve-surgery patients, the underlying disease was dominant aortic stenosis in 108 patients and aortic regurgitation in 101 patients. The remaining 94 patients underwent open-heart surgery for both aortic and mitral valve disease. Rheumatic valvular heart disease was the cause of valvular dysfunction in 285 patients (45%).

Figure 1 summarises the changes of TR severity. Before open-heart surgery, 548 patients (86%) showed TR grade <2+, and grade 2+ TR was present in 90 patients. In patients with preoperative TR grade <2+, 40 patients (7.3%) showed final TR grade ⩾3+, and nine patients (10%) showed development of significant TR in patients with grade 2+ TR (p = 0.373). The overall incidence of late significant TR was 7.7% (49/638), which was dependent on the underlying valvular lesions (fig 2). The incidence of late significant TR in patients undergoing double valve surgery was 11.7% (11/94), whereas the incidence in those undergoing mitral and aortic valve surgery was 9.6% (31/323) and 3.2% (7/221), respectively (p = 0.007).

Figure 3 shows the incidence of late significant TR according to the underlying valvular pathology and aetiology in 323 patients who underwent mitral valve surgery. Among 198 patients with predominant mitral regurgitation, the incidence of late significant TR was 15% (10/65) in rheumatic mitral regurgitation group and 5% (7/133) in the non-rheumatic (mostly myxomatous degenerative) mitral regurgitation group (p = 0.017). In the subgroup analysis, patients with rheumatic mitral regurgitation were younger than those with non-rheumatic mitral regurgitation (37 (11) vs 49 (13) years, p<0.001). At the time of mitral valve surgery, there was no significant difference in left ventricular end diastolic dimension (65 (8) vs 64 (8) mm, p = 0.193) and the peak pressure gradient of TR (34 (15) vs 31 (10) mm Hg, p = 0.20) between groups. Compared with patients with non-rheumatic mitral regurgitation, those with rheumatic mitral regurgitation showed a higher prevalence (42% vs 29%, p = 0.087) of atrial fibrillation (Af) with a larger left atrial size (58 (11) vs 54 (8) mm, p = 0.005). Mitral valve repair was performed more frequently than mitral valve replacement in both the rheumatic (71% vs 29%) and non-rheumatic mitral regurgitation group (82% vs 18%); there was no significant difference according to the aetiology of mitral regurgitation (p = 0.073).

The clinical characteristics of patients who did or did not develop late significant TR are summarised in table 1. The incidence of repeated operation to control bleeding or other complication in the acute stage was not significantly different between groups (4.6% (27/589) vs 6.1% (3/49), p = 0.495). The left ventricular ejection fraction decreased from 58 (11)% at preoperative echocardiography to 51 (12)% before hospital discharge after open-heart surgery (p<0.001). Although the left ventricular systolic dimensions decreased more significantly immediately after surgery in patients who developed late significant TR, acute changes in left ventricular diastolic dimensions and ejection fraction were not significantly different between groups. The TV annulus diameter before surgery was not significantly different between groups (31 (5) vs 32 (6) mm, p>0.5) and follow-up echocardiography showed significant increase in TV annulus in both patients who did (from 32 (6) to 40 (7) mm, p<0.001) and did not (from 31 (5) to 35 (5) mm, p<0.001) develop late significant TR. The increase in TV annulus was significantly greater in patients who developed late significant TR (3.2 (5.5) vs 8.0 (5.4) mm, p<0.001).

Af was present in 239 patients (37%) before initial surgery. Among these, 55 patients (23%) showed conversion to normal sinus rhythm at follow-up, whereas the remaining 184 patients showed persistent Af. Among 383 patients with normal sinus rhythm before surgery, Af developed in 18 patients (5%) during follow-up. Additional biatrial maze operation was done in 53 patients (22%) with Af, and 43 patients (81%) showed successful sinus conversion.

Late significant TR was found to be associated with older age, female gender, Af, a rheumatic aetiology and double valve surgery with a larger left atrium. Late significant TR was also associated with prosthesis malfunction. Multivariate analysis showed that age (hazard ration (HR), 1.0; 95% CI 1.0 to 1.1; p = 0.005), female gender (HR, 5.0; 95% CI 2.0 to 12.7; p = 0.001), rheumatic aetiology (HR, 3.8; 95% CI 1.4 to 10.3; p = 0.011), Af (HR, 2.6; 95% CI 1.1 to 6.4; p = 0.035) and a peak pressure gradient of TR at follow-up (HR, 1.1; 95% CI 1.0 to 1.1; p<0.001) were independent factors associated with the development of late significant TR.

Clinical events

Thirty-seven patients died of cardiovascular causes during follow-up including valve-related complications such as endocarditis or heart failure (n = 16) and cerebrovascular accidents (n = 14). Patients with late significant TR showed a higher mortality than those without (29/589 (4.9%) vs 8/49 (16.3%), p = 0.004). Twenty-four patients underwent repeat open-heart surgery for a variety of reasons (prosthesis malfunction (n = 8), infective endocarditis (n = 6), recurrent regurgitation after repair (n = 4), aortic dissection or aneurysm (n = 3), symptomatic severe TR (n = 2) and aggravated aortic stenosis after mitral valve surgery (n = 1)). The incidence of redo-open-heart surgery was not significantly different between patients who did and did not develop late significant TR (8.2% (4/49) vs 3.4% (20/589), p = 0.104). Patients who developed late significant TR showed a lower 8-year clinical event-free survival rate (76 (6) vs 91 (1)%, p<0.001, fig 4) compared with those without late significant TR.

DISCUSSION

In the present study, we found that significant TR developed in patients who had undergone uneventful mitral or aortic valve surgery (ie, in the absence of prosthesis malfunction or newly developed valvular lesions). Progression from mild TR to significant TR was found to be independently associated with advanced age, female gender, rhythm, rheumatic aetiology and persistent pulmonary hypertension after corrective surgery (a well-known haemodynamic variable). These factors were associated with worse clinical outcomes.

Late significant TR after left-sided valvular surgery

Although aggressive surgical procedures for moderate-to-severe TR at the initial open-heart surgery for left-sided valvular lesions are generally recommended, the challenging clinical issue is how to manage patients with mild TR at the initial surgery. To adequately address this issue requires knowledge of the real incidence of late significant TR, its clinical impact and clinical risk factors associated with progression or development of late significant TR.

The incidence of late significant TR depends on the aetiology of the underlying valvular lesions. As expected, we found that rheumatic aetiology was strongly associated with the development of late significant TR. Pancarditis as rheumatic fever progresses may have resulted in the right ventricle being more susceptible to subclinical damage and more vulnerable to progressive dysfunction and geometry change after exposure to haemodynamic changes provoked by the left-sided valvular lesions.2 Considering the higher feasibility of valve repair in degenerative mitral regurgitation, patients with rheumatic valves therefore tend to be operated on at a more advanced stage of their disease. Thus, the potential difference of the threshold or surgical indication between degenerative and rheumatic aetiologies could be indicated as a confounding factor to our observation. However, in our study, mitral valve repair was a main surgical procedure in both rheumatic and degenerative mitral regurgitation, and there was no significant difference in the frequency of mitral valve repair. Moreover, as left ventricular dimensions and development of symptom or pulmonary hypertension were our main indication of surgery for patients with mitral regurgitation, the left ventricular end-diastolic dimension and the peak pressure gradient of TR were similar between groups at the time of mitral valve surgery in this study, which abolished the possibility of different threshold or surgical indication between degenerative and rheumatic aetiologies. Therefore, we believe that our observation of a strong association of rheumatic aetiology with development of late significant TR is valid.

The gender difference in the incidence of late TR needs further clarification. A longitudinal follow-up (11.3 (8) years) study of 65 patients with rheumatic heart disease reported the incidence of late TR as 67%. That study included patients with significant preoperative organic TR and found that being female was an independent predictor of TR development.9 Thus, the present finding that female gender is an independent predictor of TR progression might be due to the confounding effect of a relatively high prevalence of mitral valve rheumatic involvement. This concept requires further investigation.

Af is the most striking risk factor associated with development of late significant TR. Matsuyama et al reported that the incidence of late significant TR during an 8.2-year follow-up of 174 patients after mitral valve surgery was 16%, and that Af was an independent predictor.11 The potential association between Af and late TR was also highlighted by a report showing the beneficial effect of a concomitant maze procedure for prevention of late TR.14 In the present study, of the 239 patients (37.5% of the total number of patients) who showed Af before surgery, 53 (22.2%) underwent a concomitant maze procedure. Interestingly, the present study did not show any beneficial effect of the maze procedure in terms of preventing late significant TR development. The inconsistency between this finding and those of other studies may reflect the different patient populations and methods of data analysis. The previous study showing a benefit of a concomitant maze procedure for prevention of late TR14 included patients with significant preoperative TR (n = 46, 27%) which required annuloplasty in up to 70% of cases (n = 32), and the prevalence of preoperative atrial fibrillation was significantly higher than in the present study (74% (126/170) vs 37% (239/638), p<0.001) with a higher rate of maze procedures (38% (48/126) vs 22% (53/239), p = 0.001). Despite these differences in baseline characteristics, we believe the present results re-emphasise the important association between Af and late TR development, since Af rather than a maze procedure is proven to be an independent risk factor. The efficacy of maze procedure in patients with markedly dilated atria, characteristic findings of rheumatic involvement, might be limited, and the exact mechanisms underlying TR progression after surgery in Af patients remain unknown. It appears that sophisticated longitudinal follow-up imaging studies examining remodelling or geographical changes in cardiac chambers are required to better address such questions.

Our results confirm previous findings that elevated pulmonary artery pressure is an important risk factor for TR progression, which support the concept that TR is the result of right ventricular hypertension and dilatation of the tricuspid annulus. Interestingly, we did not find preoperative pulmonary hypertension to be associated with TR progression, consistent with a previous study.11 The association between persistent elevation of pulmonary artery pressure at follow-up and TR progression may represent the important role of irreversible pulmonary vasculature damage at the time of surgery due to longstanding rheumatic mitral disease.9 However, the difference in TR peak pressure gradient was not large, and there was a significant overlap between patients who did and did not develop TR. This finding represents the somewhat complicated nature of late significant TR development observed in clinical practice.

Right ventricular or annular geometry including annular diameter measured by preoperative echocardiography has been reported to be associated with development of late significant TR,15 16 and we could also demonstrate that TV annulus diameter at follow-up was significantly larger in patients who developed late significant TR despite similar preoperative TV annulus diameter. However, we have failed to demonstrate the predictive value of the preoperative TV annulus size in the present study. This failure can be explained by several factors. Measurement of preoperative right ventricular dimension and TV annulus was not possible in all patients in this retrospective study, which might affect the results. However, the incidence of late significant TR was not significantly different between patients in whom the TV annulus diameter could be measured and those without TV diameter data (6.9% (31/452) vs 9.7% (18/186), p = 0.224). Thus, we believe that those patients with preoperative TV annulus diameter data can represent the total subjects of this study. Selection bias is one contributing factor. This study excluded patients with significant preoperative TR who may have a characteristic distortion of the geometry of the right ventricle or eccentricity index. Therefore, it would be very hard to characterise significant changes in geometries using conventional echocardiographic data.15^–17 Besides selection bias, marked heterogeneity in terms of methods of annular diameter measurement and the proposed cut-off values of TV annulus diameter proposed for prophylactic surgery15 16 might represent the complex geometrical alteration of TV annulus in these selected group of patients and limitation of currently available conventional imaging techniques. To overcome these technical difficulties, we believe comprehensive analysis of the data obtained using more advanced techniques, such as real-time three dimensional echocardiography,18 19 is required for clarifying the mechanisms underlying TR progression and its association with Af and other independent parameters. Moreover, as three-dimensional reconstruction has been reported to be superior to two-dimensional echocardiography for demonstration of rheumatic involvement of TV,20 a mechanistic insight of TR progression can be more easily obtained.

Limitations

We excluded five patients with preoperative grade 2+ TR who progressed by one grade even if they reached grade 3+ TR. It is well known that some medications including diuretics may reduce the haemodynamic severity of TR by relieving right ventricular distension, hence lowering right atrial pressure.21 Semiquantitative measurement of TR severity based on TR jet area was used in this study. Considering potential measurement variability or medication effects, the increase in TR grade from 2+ to 3+ in those five patients might not be a significant change in TR severity or just represent persistent TR. Moreover, patients with definite development of late significant TR, rather than persistence of moderate TR after left-sided valve surgery, were our intentional subjects for this study. Thus, we used tighter selection criteria, including a TR increase by more than one grade and final TR grade ⩾3/4 at follow-up echocardiography. We expect that the main findings of this study would not change if we added these five patients, as they comprised less than 10% of total patients with late significant TR.

Although development of signs of right heart failure is expected to occur more frequently in patients with late significant TR especially with a long follow-up duration, it was not included as a clinical event in this study. The clinical events included death, repeated open-heart surgery and hospital admission due to heart failure, which are definitely more significant events than simple signs of right heart failure, and the more frequent development of harder clinical events in this study is good enough to demonstrate the adverse clinical impact of late significant TR.

Clinical implications

The present study identified baseline risk factors (age, female gender, rheumatic aetiology and Af) associated with the progression or development of significant TR after correction of left-sided valvular heart disease. Thus, aggressive surgical intervention may be justified in such high-risk patients, even though they have mild TR. The efficacy of specific surgical procedures (the maze procedure vs tricuspid repair) warrants further investigation.