Prosthetic valve thrombosis (PVT) is a life-threatening complication. Neutrophil/lymphocyte ratio (NLR) and platelet/lymphocyte ratio (PLR) have been studied as inflammatory biomarkers in atherosclerosis, but data regarding valvular disease are lacking. The study population included patients with mitral PVT (n = 152) versus control subjects (n = 164) with functional mitral prosthesis. Transesophageal echocardiography was performed to diagnose PVT. NLR and PLR were calculated using complete blood count. C-reactive protein (CRP) levels were also analyzed. Neutrophil and platelet levels did not differ between the groups (4.9 ± 2.0 vs 4.7 ± 1.5, p = 0.84 and 254.8 ± 89.7 vs 241.5 ± 62.8 p = 0.36, respectively), but lymphocyte levels were significantly lower in patients with PVT than the controls (1.8 ± 0.7 vs 2.2 ± 0.6, p <0.001). NLR, PLR, and CRP levels were significantly higher in patients with PVT than in controls (3.2 ± 2.1 vs 2.2 ± 0.8, p <0.001; 163 ± 77.5 vs 114.9 ± 37.3, p <0.001; and 1.97 ± 3.02 vs 1.02 ± 1.22, p = 0.01, respectively). A positive correlation was observed between NLR and PLR (r = 528, p <0.001). NLR level of >2.23, measured on admission, yielded an area under the curve value of 0.659 (95% confidence interval 0.582 to 0.736, sensitivity 66%, specificity 60%, p <0.001) and PLR level of >117.78 yielded an area under the curve value of 0.707 (95% confidence interval 0.636 to 0.777, sensitivity 70%, specificity 58%, p <0.001). Multivariate analysis showed that increased PLR and inadequate anticoagulation were independent predictors of thrombosis in patients with PVT. In conclusion, patients with PVT had increased NLR, PLR, and CRP levels compared with subjects with normofunctional prosthesis, and increased PLR was an independent predictor of mitral PVT.
Prosthetic valve thrombosis (PVT) is a potentially life-threatening complication associated with high morbidity and mortality. The incidence of left-sided PVT is from 0.5% to 8% per patient-year. Inadequate anticoagulation, the early postoperative period, atrial fibrillation, left atrial (LA) enlargement, multiple valve replacement, ventricular dysfunction, and pregnancy are the major risk factors for the development of PVT. Thrombotic and inflammatory statuses of heart diseases have been evaluated using several biomarkers such as C-reactive protein (CRP), matrix metalloproteinases, pregnancy-associated plasma protein, and eosinophils. Recently, neutrophil/lymphocyte ratio (NLR) and platelet/lymphocyte ratio (PLR), which can be derived from the complete blood count, have been studied as novel markers of prognosis in patients with cardiovascular disease. These studies showed that NLR and PLR could be involved in inflammatory and thrombotic processes in patients with coronary artery disease. In contrast, data regarding the role of these parameters in valvular heart diseases is lacking. In this study, we aimed to investigate the impact of novel hematologic parameters in patients with thrombotic prosthetic mitral valves.
Methods
A total of 924 subjects with mechanical mitral prosthetic valves, who were admitted to Heart Valve Center of Koşuyolu Kartal Heart Training and Research Hospital, Istanbul, Turkey from 2009 to 2013, were analyzed retrospectively in this comparative monocentric study. The exclusion criteria included presence of LA thrombus or severe spontaneous echocontrast, treatment with fibrinolytics in the previous 24 hours, left ventricular ejection fraction (LVEF) <30%, presence of additional valve prosthesis, current therapy with corticosteroids and nonsteroidal antiinflammatory drugs, active infection, history of a systemic inflammatory process, malignancy, end-stage liver disease, renal failure, thyroid disease, and other hematologic diseases. The patients with coexisting prosthesis-related disorders such as paravalvular regurgitation were excluded as they would cause overestimation of transvalvular gradients. The patients in whom pannus and thrombus differentiation was difficult or pannus overgrowth accompanied thrombus formation were also excluded. Patients with prosthetic valve obstruction who had no thrombus, mass, or pannus in the echocardiographic study and normal prosthetic valve leaflet motion were considered a patient-prosthesis mismatch and were excluded from the study. Finally, 316 patients constituted the study population, which included patients with mitral PVT and control subjects with functional mitral prosthesis. The patient demographic characteristics, medical history, elapsed time since surgery, rhythm disorders, New York Heart Association (NYHA) functional class, incidence of thromboembolic events, and anticoagulation level at the time of admission were entered into a database. Written informed consent was obtained from the participants and the study was approved by the local ethics board.
Complete blood counts, which included total white blood cells, neutrophils, and lymphocytes, were obtained using an automatic blood counter (Cell Dyn 3700 Abbott Diagnostics Division, Santa Clara, California). NLR and PLR were calculated as the ratio of the neutrophils and platelets to lymphocytes, both obtained from the same automated blood samples that were taken at the time of admission to the study before the administration of any fibrinolytic therapy. CRP levels, which were measured using a BN 2 model nephelometer (Dade Behring, Marburg, Germany), were also entered into the database.
Transthoracic echocardiographic, 2-dimensional transesophageal echocardiographic (TEE), and real-time 3-dimensional TEE images were evaluated for each patient. The TEE studies were performed using an X7-2t transducer on an iE33 ultrasound machine (Philips Medical Systems, Andover, Massachusetts). Thrombus was recognized as a homogeneous mobile or fixed mass with similar echodensity to the myocardium located at the valve occluder and/or valve struts and was visualized in all patients with PVT by echocardiography. Differentiation of thrombus from pannus overgrowth was based on echocardiographic and clinical findings, as previously reported. The cross-sectional area and the largest diameter of the thrombus were measured on TEE images as described before. Transmitral gradients and effective orifice area were measured with 2-dimensional transesophageal echocardiography according to the current guidelines. A mean gradient of >10 mm Hg and a valve area of ≤1.5 cm 2 were indicative of mitral obstructive PVT. LVEF and LA diameter were also noted.
Continuous variables are presented as medians and interquartile ranges or as mean ± SD, as appropriate. Categorical variables are presented as observed frequencies and percentages. Group comparisons for continuous variables were tested using the Mann-Whitney U test because the data distributions were not normal. The chi-square test or Fisher’s exact test was used to compare these 3 groups for categorical variables and analysis of variance was used for continuous variables. The Bonferroni correction was applied for post hoc analysis. Multivariate logistic regression analysis was performed to identify the independent predictors of thrombosis in patients with mitral prosthesis; the variables were fitted from variables found to have marginal associations on univariate testing (p <0.05). Pearson’s correlation coefficients examined the degree of association between examined variables. The cut-off values were estimated by receiver operating characteristic curve analysis. Comparison of area under the curves (AUCs) was performed by MedCalc software program, release 12.3.0.0 (MedCalc Software, Mariakerke, Belgium). p Values <0.05 were considered significant, and the confidence interval (CI) was 95%. The statistical analyses were performed using SPSS, version 17 (SPSS, Chicago, Illinois).
Results
The study population (n = 316; women/men = 204/112) included patients with mechanical mitral PVT (n = 152) versus control subjects (n = 164) with functional mitral prosthesis. The most common valve type was the St. Jude Medical bileaflet valve (St. Jude Medical Inc., St. Paul, Minnesota). Of 152 patients with PVT, 51 patients had obstructive and 101 patients had nonobstructive PVT. The most common symptom was dyspnea. The clinical and hematologic characteristics of the study patients are presented in Table 1 . Neutrophil levels did not differ between the groups (p = 0.84) but lymphocytes were significantly lower in patients with PVT than the control subjects (p <0.001). NLR levels were significantly higher in patients with PVT compared with the controls (p <0.001). Platelet levels were higher in patients with PVT but it did not reach statistical significance between the groups (p = 0.36). However, patients with PVT had significantly higher PLR levels than the control group (p <0.001). CRP levels on admission were higher in patients with PVT in comparison with control subjects (1.97 ± 3.02 vs 1.02 ± 1.22, p = 0.01). A positive correlation was observed between NLR and PLR (r = 528, p <0.001). There was also positive correlation between NLR and CRP (r = 0.548, p <0.001) and between PLR and CRP (r = 0.528, p <0.001). Multivariate logistic regression test was employed for determining the independent predictors of PVT. The variables that were found to be significant in the univariate analysis (NLR, PLR, CRP, lymphocyte count, level of anticoagulation, and New York Heart Association class) were included in the multivariate model. Of those, PLR (odds ratio 1.011, 95% CI 1.002 to 1.019, p = 0.015) and ineffective anticoagulation (odds ratio 10.212, 95% CI 4.815 to 21.660, p <0.001) were found to be the independent predictors of PVT ( Table 2 ).
| Variable | Prosthetic Valve Thrombosis | p Value | |
|---|---|---|---|
| Yes (n = 152) | No (n = 164) | ||
| Mean age (yrs) | 48 ± 18 (16–76) | 51 ± 15 (18–77) | 0.25 |
| Women | 98 (64.5) | 106 (64.6) | 0.52 |
| Time since valve replacement (mo) | 57 ± 40 (1–264) | 61 ± 16 (1–230) | 0.72 |
| Hypertension | 26 (17.1) | 31 (18.9) | 0.87 |
| Diabetes mellitus | 12 (7.9) | 13 (7.9) | 0.98 |
| Coronary artery disease | 15 (9.8) | 16 (9.7) | 0.92 |
| Atrial fibrillation | 72 (47.3) | 78 (47.5) | 0.95 |
| New York Heart Association functional class I–II/III–IV | 88/64 (57.9/42.1) | 132/32 (80.5/19.5) | <0.01 |
| Adequate anticoagulation | 24 (15.8) | 147 (89.6) | <0.01 |
| Hemoglobin (g/dl) | 12.4 ± 1.8 | 12.3 ± 1.6 | 0.84 |
| Red cell distribution width (%) | 17.1 ± 0.4 | 16.9 ± 0.4 | 0.36 |
| Platelet count (103 μl) | 254.8 ± 89.7 | 241.5 ± 62.8 | 0.36 |
| Mean platelet volume (fl) | 10.3 ± 21.1 | 8.5 ± 1.5 | 0.31 |
| White blood cell count, (×109) | 6.8 ± 1.5 | 7.2 ± 1.1 | 0.68 |
| Neutrophils (×109) | 4.9 ± 2.0 | 4.7 ± 1.5 | 0.84 |
| Lymphocytes (1,000/μl) | 1.8 ± 0.7 | 2.2 ± 0.6 | <0.001 |
| Neutrophil/lymphocyte ratio | 3.2 ± 2.1 | 2.2 ± 0.8 | <0.001 |
| Platelet/lymphocyte ratio | 163 ± 77.5 | 114.9 ± 37.3 | <0.001 |
| Variable | Multivariate p Value | Multivariate OR and 95% CI |
|---|---|---|
| NLR | 0.544 | 1.147 (0.736–1.788) |
| PLR | 0.015 | 1.011 (1.002–1.019) |
| C-reactive protein | 0.263 | 1.217 (0.863–1.715) |
| Lymphocyte count | 0.133 | 0.562 (0.265–1.192) |
| Level of anticoagulation | <0.001 | 10.212 (4.815–21.660) |
| New York Heart Association class | 0.150 | 0.550 (0.244–1.241) |
Receiver operating characteristic curve analysis was performed to detect the best cut-off value of NLR and PLR in the prediction of thrombosis in patients with prosthetic valves. NLR level of >2.23, measured on admission, yielded an AUC value of 0.659 (CI 95% 0.582 to 0.736, sensitivity 66%, specificity 60%, positive predictive value 62%, negative predictive value 64%, p <0.001), and PLR level of >117.78 yielded an AUC value of 0.707 (CI 95% 0.636 to 0.777, sensitivity 70%, specificity 58%, positive predictive value 61%, negative predictive value 67%, p <0.001). To identify the most accurate parameter, we compared both AUC of receiver operating characteristic curves and no significant difference was detected between NLR and PLR (p = 0.247; Figure 1 ).
The echocardiographic parameters of PVT subgroups of patients (obstructive vs nonobstructive) and control subjects are listed in Table 3 . No significant difference was observed between these subgroups in terms of LVEF and LA diameter. From analysis of variance, there was significant difference among the 3 groups with regard to the mitral valve area (p <0.001) and mean transvalvular gradient (p <0.001); the patients with obstructive PVT had lower valve area and higher mean transvalvular gradient compared with nonobstructive PVT subgroup and control subjects ( Table 3 ). The obstructive subgroup had significantly higher NLR than nonobstructive PVT subgroup and control subjects (3.85 ± 1.91 vs 2.84 ± 2.18, p = 0.004 and 3.85 ± 1.91 vs 2.25 ± 0.85, p <0.001, respectively). In contrast, both obstructive and nonobstructive PVT subgroups had significantly higher PLR than control group (163.6 ± 75.4 vs 114.9 ± 37.3, p <0.001 and 161.9 ± 82.3 vs 114.9 ± 37.3, p <0.001, respectively). Similar to NLR, CRP levels were significantly higher in obstructive subgroup than in nonobstructive PVT subgroup and control subjects (2.98 ± 4.2 vs 1.46 ± 2.0, p = 0.002 and 2.98 ± 4.2 vs 1.0 ± 1.22, p <0.001, respectively). There was a modest-to-strong positive correlation between the mean transvalvular gradient and NLR (r = 0.693, p <0.001). Furthermore, there was a mild but significant correlation between the mean thrombus area and NLR (r = 0.345, p <0.001). In contrast, there was a mild correlation between PLR and the mean transvalvular gradient (r = 0.239, p <0.001) and between PLR and the mean thrombus area (r = 0.285, p <0.001). There was also mild but significant correlation between CRP and mean transvalvular gradient and between CRP and mean thrombus area (r = 189, p = 0.005 and r = 203, p <0.001, respectively).
