Background

Mitral valve replacement (MVR) in children is a rare and challenging procedure with reported high complication rates [1, 2]. The surgical challenge of MVR in children arises because of the small annular diameter, the associated lesions that may affect valvular morphology and annular growth, and the unavailability of suitable mitral valve prostheses [3,4,5]. Mitral valve repair in children offers a durable option, with long-term freedom from mitral valve reinterventions; therefore, MVR is performed when the repair cannot be performed or is inadequate [6]. The availability of small prosthetic mitral valves and advances in cardiopulmonary bypass techniques in children have led to an increase in the number of mechanical mitral valves in young children [7,8,9]. The use of a small mitral valve prosthesis in children may increase the risk of heart block and valve-related thrombosis. Furthermore, it can affect long-term survival, including the need for redo surgery to upsize the mitral valve [1, 10, 11]. Research is ongoing to develop dynamic valves that expand with annular growth [12]; however, the durability of biological materials in children, especially on the left side, is unclear. Mechanical mitral valves could provide a reasonable solution in these patients because of their long-term durability; however, the need for anticoagulation in the young population may increase the risk profile of mechanical valves [13]. Several factors can jeopardize the long-term outcomes of MVR. Worse survival was reported with the supra-annular valve position [2] and with a small valve size [14]; therefore, tailoring surgical techniques to avoid small valves was associated with improved outcomes [14]. Recent studies have shown that MVR in children is associated with low morbidity and mortality, especially with new mitral valve prostheses [13, 15,16,17,18].

Studies evaluating the outcomes of MVR in children younger than 2 years are scarce, and the outcomes could be further affected by the use of a smaller prosthesis, the need for anticoagulation, and the need for mitral valve reoperations [13]. Therefore, the goal of this study was to report the outcomes of mechanical MVR in children under 2 years old and to evaluate factors affecting long-term valve-related thrombosis, mitral valve reoperation, and survival.

Methods

Patients

This retrospective cohort study included children under 2 years of age who underwent MVR at a single center from 2000 to 2023. The study included patients who underwent mechanical MVR. Patients who were older than 2 years, had univentricular hearts, had biological valves, or had concomitant aortic or pulmonary valve replacement were excluded. The study was approved by the hospital’s institutional review board.

Study data and outcomes

The data collected included age in months at the time of the indexed operation, sex, weight in kg, cardiac diagnosis, the presence of mitral valve lesions, and history of previous cardiac surgery. Reported operative data included mitral valve sizes, types, and positions (annular vs. supra-annular). Postoperative complications included renal failure, sepsis, complete heart block (CHB) requiring permanent pacemaker (PPM) insertion, left ventricular outflow tract obstruction, extracorporeal membrane oxygenation insertion, intensive care unit (ICU) and hospital stay, and hospital mortality. The follow-up outcomes included bleeding, valve-related thrombosis, mitral valve reoperation, and survival.

Anticoagulation protocol

Heparin infusion was started 24 h postoperatively along with warfarin until the target international normalization ratio (INR) of 2.5–3.5 was achieved. Patients were followed up after discharge at the outpatient clinic. Low-molecular-weight heparin was used for patients who were not tolerant or noncompliant with warfarin or who were at high risk of bleeding. Additional antiplatelet therapy with aspirin was used in selected patients according to the discretion of the treating physicians.

Surgical techniques

All patients underwent mitral valve surgery through median sternotomy and cardiopulmonary bypass. The mitral valve was approached via left or right atriotomies depending on the concomitant pathology and the surgeons’ preference. The mitral valve was inserted into the annular or supra-annular position.

Follow-up protocol

The patients were followed for 1 week after discharge at the anticoagulation clinic and postoperative clinic. The frequency of follow-up in the anticoagulation clinic was determined according to the INR. For patients with a stable INR, follow-up was scheduled every 4 weeks.

Statistical analysis

The normality of the distribution of continuous variables was evaluated, and normally distributed data are presented as the mean and standard deviation, while nonnormally distributed data are expressed as the median (25th and 75th quartiles). Categorical data are presented as numbers and percentages. Time-to-event data are presented as Kaplan–Meier curves, and the curves were compared between annular and supra-anular positions using the log-rank test. Factors affecting time-to-event outcomes were evaluated with univariate Cox regression analysis, and the hazard ratio (HR) and its confidence interval were reported. Analyses were performed using Stata 17 (Stata Corp, College Station, TX, USA).

Results

Baseline and operative data

The study included 23 patients who underwent mechanical MVR; their ages ranged from 2 to 24 months. The mean weight was 7.6 ± 2.9 kg, and 11 (48%) patients were males. The most common diagnosis was congenital MR (n = 6; 26%). Mitral regurgitation was the most common mitral valve lesion (n = 17; 74%), two patients had mitral stenosis (9%), and four had mixed lesions (17%).

The size of the mitral valve prosthesis ranged from 16 to 25 mm. The most common types were St. Jude (n = 15) (St. Jude Medical Inc.; Minneapolis, MN, USA), and CarboMedics valve (n = 4) (CarboMedics, Inc.; Austin, Texas). In 6 patients (26%), the mitral valve was inserted in a supra-annular position, and in the remaining 17, the mitral valve was inserted in the annular position (Table 1).

Table 1 Baseline and operative data in children younger than 2 years who underwent mechanical mitral valve replacement
Full size table

Postoperative outcomes

Renal failure occurred in three patients (13%), sepsis in four patients (17%), and complete heart block in seven patients (32%). There was no significant difference in PPM insertion between annular (n = 5, 31%) and supra-annular (n = 2, 33%) valves (p > 0.99). The median duration of ICU stay was 14 days (range 7–25), and the median duration of hospital stay was 27.5 days (range 14–60). Operative mortality occurred in three patients (13%). Warfarin was used in 17 patients (Table 2).

Table 2 Postoperative outcomes in children younger than 2 years who underwent mechanical mitral valve replacement
Full size table

Follow-up events

Bleeding

Bleeding was reported in five (22%) patients; four were managed conservatively, but one had major intracranial hemorrhage (ICH) requiring craniotomy. Three of the patients who had bleeding were on warfarin, and two were not (p = 0.58). The sites of bleeding were dental (n = 2), mouth (n = 1), pulmonary (n = 1), and ICH (n = 1).

Valve thrombosis

Valve thrombosis was defined as a 50% increase in the mean mitral inflow gradient by echocardiography or a stuck mitral valve leaflet on fluoroscopy or echocardiography. Nine patients had thrombosis (39%) at least once (n = 4), twice (n = 2), five times (n = 2), or nine times (n = 1). Alteplase was used in 7 patients on 26 occasions [once (n = 2), twice (n = 2), five times (n = 2), and nine times (n = 1)], and it was successful in treating the thrombus in 24 instances. In two events, the patients presented too late with severe pulmonary edema and died. Two patients were managed by redo surgery: one with removal of a clot and one with redo mitral valve replacement.

The percentages of patients who were free from thrombosis at 1, 2, and 5 years were 67%, 55%, and 55%, respectively. Four out of the 6 patients with supra-annular valves had thrombosis, and 5 out of the 17 patients with annular valves had thrombosis. The rates of freedom from valve-related thrombosis in the annular position were 86%, 71%, and 53% after 1, 2, and 5 years, respectively, while they were 50% in the supra-annular position after 6 months (p < 0.001) (Fig. 1).

Fig. 1
figure 1

Freedom from valve-related thrombosis in patients with annular and supra-annular mitral valves

Full size image

Thrombosis was not related to age [HR: 0.95 (95% CI: 0.84–1.07); p = 0.381], sex [HR: 0.99 (95% CI: 0.26–3.71); p = 0.988], weight [HR: 1.06 (95% CI: 0.85–1.33); p = 0.607], or valve size [HR: 0.92 (0.72–1.17); p = 0.487]. The incidence of valve thrombosis was significantly greater in patients who were not on warfarin (p < 0.001) (Fig. 2).

Fig. 2
figure 2

Freedom from valve-related thrombosis in patients treated with and without warfarin

Full size image

Four patients required postoperative extracorporeal membrane oxygenation (ECMO) support, two of whom underwent additional repair of the atrioventricular canal, and one patient underwent additional tetralogy of Fallot repair. Mortality occurred in two patients with ECMO support [19].

Mitral valve reoperation

Seven patients underwent redo MVR, two of whom underwent reoperations twice. The median time to reoperation was 19 months (range 4–45). The second redos occurred after 121 and 130 months. One patient had a redo with a tissue valve, while the others had mechanical valves. The rate of freedom from valve reoperation was 76% after 1 year, 69% after 2 years, and 57% after 5 years. Four patients had annular valves, and three patients had supra-annular valves. The rates of freedom from mitral valve reoperation at 1, 2, and 5 years were 92%, 84%, and 70%, respectively, for annular valves, and they were zero% after 7 months in the supra-annular position (p < 0.001) (Fig. 3).

Fig. 3
figure 3

Freedom from mitral valve reoperation in patients with annular and supra-annular valves

Full size image

Redo-MVR was more common in younger patients [HR: 0.84 (95% CI: 0.72–0.98); p = 0.029], while it was not related to sex [HR: 0.83 (95% CI: 0.18–3.82); p = 0.814], weight [HR: 0.90 (95% CI: 0.69–1.17); p = 0.420], or valve size [HR: 0.73 (95% CI: 0.49–1.09); p = 0.121].

Survival

The median follow-up duration was 26 months (range 4–126). Six deaths (26%) were reported, with thrombosis being the direct cause of death in two patients: three died because of sepsis, and one died because of pulmonary hypertension crisis. Survival was 78% after 1 year and 73% after 5 years. Survival at 5 years was 88% in the annular position, 50% after 3 months, and 25% after 14 months in the supra-annular position (p = 0.009) (Fig. 4).

Fig. 4
figure 4

Survival in patients with annular and supra-annular mitral valves

Full size image

Survival was not related to age [HR: 0.78 (95% CI: 0.60–1.004); p = 0.054], sex [HR: 0.99 (95% CI: 0.20–4.96); p = 0.994], weight [HR: 0.90 (95% CI: 0.67–1.20); p = 0.464], or valve size [HR: 0.82 (95% CI: 0.57–1.18); p = 0.275].

Discussion

Cardiac valve defects are common congenital cardiac lesions in children, and repair is the procedure of choice for restoring long-term cardiac function [20]. The optimal mitral valve prosthesis in children is still a subject of ongoing research, and the results of the available prostheses remain suboptimal. Mechanical valves offer a durable solution but have a high risk of thromboembolic complications and lifelong anticoagulation therapy [21]. On the other hand, tissue valves are prone to early degeneration when implemented in children [22], and both valve types have fixed diameters that do not grow with the patients. A small valve size predicts worse outcomes after MVR in children, and several techniques have been adapted for inserting larger prostheses that might persist longer [14]. The long-term outcomes after MVR in children younger than 2 years remain a concern. This study evaluated the outcomes after mechanical MVR in children younger than 2 years and assessed possible factors that might affect the outcomes. The study included 23 patients whose mitral valve size ranged between 16 and 25 mm, and 6 patients had a supra-annular valve. The most common postoperative complication was a complete heart block requiring PPM insertion (32%), and operative mortality was reported in three patients (13%). Seven patients had CHB, two of whom had AV canals, one who had VSD repair, and one who had TOF repair. Valve-related thrombosis was a common complication, and 28 attacks were reported in 9 patients, two of whom underwent surgical interventions. Thrombosis was more common in patients not on warfarin and with a supra-annular valve. Additionally, patients with supra-annular valves had more mitral valve reoperations and lower survival.

Supra-annular valve implantation has been described in children with small native annular diameters [23]. The technique was theoretically introduced to expand the life of the valve; however, Barker and associates reported that the use of the supra-annular mitral valve did not extend the time to the second replacement, and on the contrary, it affected hemodynamics [24]. Kanter and colleagues evaluated 15 children with supra-annular MVR and reported permanent pacemaker insertion in 3 patients and early thrombosis in 3 patients. On follow-up, six patients underwent mitral valve reoperations for pannus formation or thrombus [23]. Giordano and coworkers studied seven children with supra-annular MVR and reported reoperation in two patients for thrombus or pannus formation [25]. Tierney and colleagues reported lower survival in patients with supra-annular MVR, similar to our results, but with a lower rate of complete heart block [2]. Kwon and associates compared supra-annular MVR using polytetrafluoroethylene grafts (n = 8) to annular MVR (n = 33) in children < 20 kg. They did not report differences in survival between the groups [26].

Anticoagulation protocols for mechanical valves in children vary widely between centers [27]. In this study, patients were treated with different anticoagulation regimens after MVR. The main anticoagulation protocol used warfarin to achieve a target INR between 2.5 and 3.5. However, the use of warfarin is challenging in young patients, and some patients require the use of low-molecular-weight or unfractionated heparin. In one study, the durability of mechanical valves in children was affected by the frequent occurrence of valve-related thrombosis, making the reoperation rate comparable to that of biological valves [28]. This study showed that valve-related thrombosis was more common in patients who were not receiving warfarin therapy. The anticoagulation regimen should be tailored to the patient’s characteristics and risk factors. Shin and colleagues reported that the rate of freedom from anticoagulation-related complications (bleeding/thrombosis) was 94% at 5 years after mechanical valve replacement in children [29]. Most of our patients with valve-related thrombosis were managed with alteplase, with successful outcomes. Surgery was required for only two patients with valve-related thrombosis. A previous study reported that alteplase was successful in relieving prosthetic mitral valve thrombosis in 19/20 children [30].

The results of this study demonstrated that mechanical MVR in children younger than 2 years was associated with increased morbidity and mortality. The durability of valves in children is a major concern. A recent study tested the potential use of balloon-expandable heart valves that can adapt to annular growth in children [12]. However, the issue of accelerated bioprosthetic degeneration in children remains challenging. These results indicate that the optimal valve prosthesis in children has not yet been identified, and further research is required to improve the durability and decrease the complication rate in pediatric patients with artificial heart valves.

Limitations

This study has several limitations. First, the retrospective design has inherent biases. Several factors could affect the outcomes and were not thoroughly evaluated, and the choice of technique or prosthesis could be affected by surgeons’ preferences and experience. However, MVR in children younger than 2 years is rare, and prospective randomized studies are not practical. Second, the study evaluated mechanical MVR; no alternative options were compared to this valve type. Third, the study is limited by the sample size, but this is attributed to the low frequency of the procedure. Finally, this was a single-center study, and the protocols used to select patients for surgery and anticoagulation regimens might not be similar to those used in other centers.

Conclusions

Mechanical MVR in children younger than 2 years is associated with high complication rates, especially in patients with supra-annular valves, including thrombosis and reoperation. Anticoagulation with warfarin remains challenging in infants. On the other hand, not using warfarin appears to be associated with an increased risk of valve-related thrombosis. However, further studies evaluating alternative options are needed.