Results
Between January 2010 and February 2020, 556 female patients underwent mitral valve surgery at our center, of whom 355 (64%) were subjected to MVR and were included in the study, and 201 (36%) were subjected to mitral valve repair. Among those subjected to MVR, 174 (49%) received MBP (Group I) and 181 (51%) received MMP (Group II). In the overall included population, 266 (75%) patients were in child- bearing period [Group I: 134(77%); Group II: 132 (73%)]. Moreover, 263 (74%) patients had rheumatic mitral valve disease [243 (74%) severe mitral regurgitation grade (≥V); 220 (62%) annular dilatation], and 128 (36%) had AF.
Comparison of preoperative patient demographics, socioeconomic status, mitral valve pathology, and echocardiographic data showed no statistically significant differences between groups (Table 1). Operative parameters were also not significantly different between groups, except cardiopulmonary bypass time that was significantly shorter in Group I than in Group II (p  = 0.0001; Table 2). Moreover, the postoperative ICU and clinical follow-up course were better and hospital and ICU stays were significantly shorter in Group I than in Group II (p< 0.0001) (Table 3), probably because patients in Group II needed to wait for the target therapeutic international normalized ratio (INR) (range 2.5-3 folds the normal value) and had more postoperative complications than those than those in Group I.
Postoperative NYHA-class status was much improved in Group I compared to that in Group II (p  = 0.04). Notably, the frequency of mitral valve re-replacement was higher in Group II than in Group I (3.7% vs. 0.6 %; p  = 0.02).
Postoperative complications, including wound infection, pericardial tamponade, AF, pneumothorax, pneumonia, and ARF, showed no significant differences between groups. Immediate operative death was only seen in 0.6% of the patients in Group II. After 10years of follow-up, the postoperative transvalvular pressure gradient was significantly lower and the vena contracta was significantly wider in Group I than in Group II (p  < 0.0001and  = 0.006, respectively).
At 30days postoperatively, 3 (0.8%) patients in the overall population died [one in Group I (0.6%) and two in Group II (1.2%)]. The most common complication at 30 days after surgery was AF [68 patients (19%)].Moreover, 36 (10%) patients experienced pneumonia, 5 (1.4%) wound infection, 28 (7.9%) pericardial tamponade, 13 (3.7%) pneumothorax, 9 (2.5%) pleural effusion, and 14 (3.9%) stroke. Additionally, 34 (9.6%) patients had ARF, with only 10 (2.8%) requiring temporary hemodialysis, while 29 (8.2%) patients suffered from A-V block grade III, and 27 (7.6%) had pacemaker implantation. Re-exploration for postoperative bleeding occurred in 22 (6.2%) patients and was significantly more frequent among Group II [18 (10%)] than in Group I [4 (2.4%)] patients (p < 0.0001). Abortion occurred in 32 (9%) women in child-bearing period and was significantly more frequent among Group II [27 (15%)] than in Group I [5 (3%)] patients (p = 0.0001). Anticoagulant complication-free pregnancy occurred in 242 (68.2%) women in child-bearing period, and the rate was significantly higher in Group I [141 patients (81%)] than in Group II [101 (56%)] (p < 0.0001; Table 4).
Oral anticoagulants such as coumadin were available at all times for all patients, but compliance to these drugs was 81%. The INR of Group II patient was in the therapeutic target most of the times. Moreover, 5 out of 7 patients in this group who had mitral valve re-replacement due to poor oral anticoagulant compliance or failure of coumadin treatment to reach the INR therapeutic target in pregnant woman resulting in life-threatening mechanical valve thrombosis in a pregnant woman who received MMP (5 ± 2) years before pregnancy. However, all a majority of them were successfully treated with MBP re-replacement but had a significant risk of miscarriage. In contrast, there was only one case for mitral valve re-replacement in Group I due to structural valve deterioration. Additionally, 18 (10%) Group II patients aborted due to teratogenicity related to coumadin, representing about 67% (18/27) of all abortion cases in this group.
The Kaplan–Meier curves of postoperative complications and survival rate over the 10-year follow-up period are presented in Figure 1. The survival rate was significantly higher in Group I than in Group II (88%vs 63% p  < 0.0001).The 95% confidence interval of Cox hazards survival regression ratio was significantly different between groups [0.1926 (0.0759–0.4889), p = 0.0001], whereas the adjusted 95% confidence interval for preoperative variables (age, logistic EuroScore-I, mitral regurgitation grade 4) was not [0.5581(0.3254–0.9581),  = 0.059670].