Results
The search strategy and removal of duplicates retrieved 850 title-abstracts for review. Of these, 179 full-text articles were eligible for assessment. Sixty-three articles did not meet the proposed inclusion criteria (most often due to heterogeneity of mitral valve disease or presence of symptoms in the study cohort) and 102 articles were reviews or editorials (Figure 1). Three retrospective observational cohort studies with propensity score adjustment analyses were included for data extraction and analysis. Two additional cohort studies that only investigated a watchful waiting strategy were included to better understand heterogeneity in results of the conservative approach. Two studies from Kang and colleagues were derived from the same database. Hence, the study encompassing the larger time frame and that which included an additional study center's data was included.
Table 1 summarizes the characteristics of the included studies and provides insight into potential biases in each study. The definition of early surgery varied between studies, and was performed within 3 to 6 months upon diagnosis of severe MR. Additionally, only the studies by Rosenhek and colleagues and Kang and colleagues specifically excluded patients with class II triggers (atrial fibrillation or pulmonary hypertension) on study entry. However, the remaining studies do report results of subgroup analyses performed in patients without class II triggers. Also of note, patients were younger in the studies by Rosenhek and colleagues, and Kang and colleagues.
Long-term All-cause Mortality
Two of the three studies reported a significant difference in all-cause mortality between the early surgery and watchful waiting groups (Table 2). In the three studies with a comparator arm, 10- to 12-year survival was 89% to 91% in the early surgery cohort and 62% to 88% in the conservative cohort. However, in an isolated analysis of the watchful waiting approach, Rosenhek and colleagues reported an 8-year survival rate (91%), equivalent to that of the survival of early surgery groups.
Pooled analysis of the overall study cohorts revealed a significant reduction in long-term mortality with an early surgery approach [HR =0.46; 95% confidence interval (CI): 0.24–0.88]. This survival benefit was even more pronounced in a pooled analysis of the propensity score matched cohorts (HR =0.38; 95% CI: 0.21–0.71) (Figure 2). Furthermore, the benefit of early surgery on survival persisted after a sensitivity analysis was performed to address the significant heterogeneity between studies (I=67.13%, P=0.05 for Q statistic). Given the strong effect size of the three included studies, 21 unpublished null result studies would be required to bring the new pooled P value to a non-significant level (fail safe N test). To determine whether even earlier surgical intervention improved survival, a pooled analysis of the subgroup without atrial fibrillation or pulmonary hypertension was performed. The reduction in all-cause mortality persisted when those without class II triggers underwent early surgery compared with watchful waiting (RR =0.85; 95% CI: 0.75–0.98). However, significant heterogeneity was noted (I=86.93%, P<0.001 for Q statistic).
(Enlarge Image)
Figure 2.
Pooled analysis of all-cause mortality stratified by timing of surgical intervention.
Operative Mortality
All studies (including the two without an early surgery group) reported operative mortality rates ≤1% (Table 2). Aggregate mortality rates across all five studies revealed an operative mortality of 0.7% in the early surgery cohorts and 0.7% in the watchful waiting cohorts. Meta-analysis corroborated the absence of an incremental risk of operative mortality with a watchful waiting approach (RR =0.55; 95% CI: 0.13–2.32) (Figure 3).
(Enlarge Image)
Figure 3.
Pooled analysis of operative mortality stratified by timing of surgical intervention.
Repair Rate
In two of the three comparative effectiveness studies, the rate of mitral valve repair compared with replacement was significantly higher in the early surgery cohorts (Table 2). In the third study, Montant and colleagues achieved a 100% repair rate in both groups. However, exclusive of the study by Montant and colleagues, the mitral valve was surgically repaired in less than 95% of cases in all studies, regardless of treatment cohort.
Pooled analysis of the comparative effectiveness studies demonstrated that early surgery significantly increased the likelihood of mitral valve repair compared with watchful waiting (RR =1.10; 95% CI: 1.02–1.18) (Figure 4). Although excluded from the pooled analysis because they lack a comparator arm, the watchful waiting studies by Rosenhek and colleagues and Enriquez-Sarano and colleagues both reported repair rates (82.9% and 90.9%, respectively) less than that of the early surgery cohorts in the comparison studies (93% to 100%).
(Enlarge Image)
Figure 4.
Pooled analysis of mitral valve repair rate stratified by timing of surgical intervention. Note that the study by Montant et al. could not be included because there were zero mitral valve replacements in both groups.
Repeat Mitral Valve Surgery
Three studies included data on the incidence of repeat mitral valve surgery. Kang and colleagues reported similar rates of repeat mitral valve surgery in the early surgery and watchful waiting groups (1.7% vs. 1.0%, respectively, P=1.0). Rates of reoperation in the other two studies were higher; 5.6% of patients required reoperation in the early surgery cohort of Montant and colleagues' study, and mitral reoperations were performed in 5.7% of patients in the watchful waiting cohort of Rosenhek and colleagues' study. Because only one study included data from both treatment arms, pooled analyses of the incidence of repeat mitral valve surgery were not performed.
Atrial Fibrillation
Three studies analyzed the incidence of new atrial fibrillation. Although more patients in the early surgery cohort developed atrial fibrillation within the first 3 months of study entry, the overall incidence of atrial fibrillation did not significantly differ between treatment groups over long-term follow-up (24.7% early surgery vs. 27.0% watchful waiting, P=0.89). The incidence of atrial fibrillation was lower in the watchful waiting cohort in the study by Kang and colleagues (21.2% early surgery vs. 9.3% watchful waiting, P<0.0001). This lower rate of atrial fibrillation closely approximated that reported by Rosenhek and colleagues (8.5%). Meta-analyses were not performed because only two studies contained data for both treatment arms.
Development of Class I or II Triggers
Only the Rosenhek series automatically referred patients for surgery when class I or II triggers developed. In their series, 55% of patients remained free of surgical triggers 8 years after study inclusion. In another prospective study of asymptomatic patients with degenerative MR, only 46% of patients remained free from surgical intervention at 5 years. However, it should be noted that physician or patient preference was the primary surgical indication in 47 of the 232 patients that underwent surgery.