Decrease in blood platelet count has been described after percutaneous coronary intervention and surgical valve replacement, although no study has been performed in the setting of transcatheter aortic valve implantation (TAVI). The aim of this study was to address the incidence, mechanism, and impact of blood platelet count decrease after TAVI. One hundred forty-four consecutive patients (mean age 84 ± 7 years, 64 men) with severe symptomatic aortic stenosis who underwent TAVI from December 2007 to July 2011 were enrolled. Blood platelet count was recorded before and after aortic valve implantation. Decrease in blood platelet count was compared with in-hospital major adverse cardiovascular events (death, stroke, and major or life-threatening bleeding). Blood platelet count decreases occurred in all but 1 patient. The percentage of platelet count decrease averaged 34 ± 15% and was 24% greater than blood protein decrease. Decrease in platelet count was associated with a higher rate of prosthesis migration, longer x-ray and procedural times, and larger contrast amounts (230 ± 128 ml for the third tertile vs 170 ± 77 ml for the second and first tertiles, p = 0.0006), but no association was observed with regard to changes in bilirubin. In-hospital major adverse cardiovascular events (n = 50 [35%]) were observed more frequently in patients with severe platelet count decreases (21% for the first tertile, 35% for the second tertile, and 48% for the third tertile, p = 0.02). Finally, the percentage of blood platelet count decrease was the only predictor of in-hospital major adverse cardiovascular events (odds ratio 1.67, 95% confidence interval 1.05 to 2.67, p = 0.03). In conclusion, a decrease in platelet count is a common phenomenon after TAVI, and its severity is associated with poor outcomes.
In recent years, transcatheter aortic valve implantation (TAVI) using stent-based prostheses has become an attractive alternative for high-risk elderly patients with symptomatic aortic stenosis. This technique provides short- and long-term clinical and hemodynamic improvements. The most frequent complications associated with TAVI are bleeding and stroke. Decreases in platelet counts have been observed after surgical aortic valve replacement and also occur after TAVI in clinical practice. However, a decrease in platelet count and its clinical impact on outcome after TAVI have never been studied.
Methods
This study included 144 consecutive patients who underwent TAVI with self-expandable valves using the transfemoral approach from December 2007 to July 2011 at Henri Mondor University Hospital. The definition of severe aortic stenosis was determined by the echocardiographic findings of an aortic valve area <0.8 cm 2 or 0.6 cm 2 /m 2 , a peak aortic jet velocity >4.0 m/s, or a mean aortic valve gradient >40 mm Hg. All patients were screened before TAVI to determine whether they were considered unsuitable for surgical aortic valve replacement, according to a consensus between cardiac surgeons and cardiologists. Only patients with native aortic stenoses were included in the study. All patients provided written informed consent before enrollment in the registry.
The TAVI procedure was previously reported in detail. Valve implantation was performed using the retrograde approach. The Medtronic CoreValve (Medtronic, Inc., Minneapolis, Minnesota) was used for all patients. Vascular access and closure were performed by means of a suture device (Prostar XL; Abbott Vascular, Redwood City, California). Dual-antiplatelet treatment with aspirin 75 mg and clopidogrel 75 mg was started the day before the procedure and followed thereafter, except for patients requiring oral anticoagulation by vitamin K antagonist. During the procedure, unfractionated heparin was injected to maintain an activated coagulation time of >250 seconds.
Blood platelet count was measured before valve implantation (baseline) and every day after the procedure until discharge. Blood samples were drawn by venipuncture every morning. Baseline and nadir platelet count after TAVI was used to determine the maximum percentage of platelet count decrease. Blood was collected into ethylenediaminetetraacetic acid Vacuette tube (Greiner Bio-One, Kremsmünster, Austria) and processed for automatic platelet numeration (Beckman Coulter, Brea, California). The normal range of platelet count values for the automatic numeration is 150 × 10 9 /L to 500 × 10 9 /L, and the coefficient variation is <5%. The nadir of platelet count was defined as the minimum platelet count before any blood transfusion during the hospitalization period.
Clinical follow-up was carried out through clinical visits or phone calls during the hospital stay and after 30 days. Major adverse cardiovascular events (MACEs) were prospectively collected during this period. The primary outcome was in-hospital MACEs, defined as all-cause death, life-threatening and major bleeding, and stroke during the hospitalization period as defined by the Valve Academic Research Consortium classification. Device success and 30-day combined safety data were evaluated according to the Valve Academic Research Consortium criteria. The combined safety end point was defined as follows: all-cause mortality, major stroke, life-threatening bleeding, stage 3 acute kidney injury, periprocedural myocardial infarction, major vascular complications, and a repeat procedure for valve-related dysfunction.
Continuous variables with normal distributions are expressed as mean ± SD and nominal variables as percentages. To compare numerical data between groups, paired and unpaired Student’s t tests were used as appropriate. Nominal variables were compared using chi-square tests. Kendall’s correlation was used for trend testing. Multivariate analyses were performed using linear regression. Survival time-to-event analysis was performed using Kaplan-Meier curves. Two-tailed p values <0.05 were considered as statistically significant.
Results
Patients’ baseline characteristics are listed in Table 1 . All patients had severe symptomatic aortic stenosis, with a mean peak aortic jet velocity of 4.3 ± 0.8 m/s, a mean aortic gradient of 47 ± 17 mm Hg, and a mean aortic valve area of 0.7 ± 0.2 cm 2 . After aortic valve implantation, the mean peak aortic jet velocity and mean aortic gradient decreased to 2.3 ± 2.3 and 9 ± 4 mm Hg, respectively, and the mean aortic valve area increased to 2.0 ± 0.4 cm 2 . Significant postprocedural aortic regurgitation (mild to severe, grade ≥2/4) was observed in 26 patients (18%).
| Variable | Value |
|---|---|
| Age (yrs) | 84 ± 7 |
| Men | 64 (44%) |
| European System for Cardiac Operative Risk Evaluation score (%) | 24 ± 12 |
| Society of Thoracic Surgeons score (%) | 12 ± 8 |
| Aortic maximal velocity (m/s) | 4.3 ± 0.8 |
| Aortic valve area (cm 2 ) | 0.7 ± 0.2 |
| Left ventricular ejection fraction (%) | 49 ± 14 |
| Diabetes mellitus | 35 (24%) |
| Hypertension ∗ | 109 (76%) |
| Dyslipidemia † | 78 (54%) |
| Previous myocardial infarction | 29 (13%) |
| Previous vascular disease | 29 (20%) |
| Previous cerebrovascular event | 17 (12%) |
| Previous cardiac surgery | 20 (14%) |
| Chronic obstructive pulmonary disease | 37 (26%) |
| New York Heart Association class | 2.8 ± 0.6 |
| Creatinine (μmol/L) | 111 ± 38 |
| Estimated glomerular filtration rate (ml/min) | 55 ± 23 |
| Hemoglobin (g/dl) | 11.9 ± 1.7 |
| Platelet count (10 9 /L) | 216 ± 67 |
∗ Blood pressure >140/90 mm Hg, previous diagnosis of hypertension, or use of antihypertensive medication.
† Low-density lipoprotein >100 mg/dl or use of lipid-lowering medication.
Blood platelet count at baseline before TAVI averaged 216 × 10 9 /L ± 67 × 10 9 /L (range 77 × 10 9 /L to 441 × 10 9 /L). Decrease in platelet count after TAVI occurred in all but 1 patient. The minimum platelet count averaged 170 × 10 9 /L ± 54 × 10 9 /L (range 41 × 10 9 /L to 336 × 10 9 /L) and was observed 2.5 ± 1.1 days after TAVI ( Figure 1 ). The decrease in platelet count averaged 34 ± 15% (18 ± 7% for the first tertile, 33 ± 4% for the second tertile, and 55 ± 10% for the third tertile). Overall, 90 patients had moderate thrombopenia (platelet count 50 × 10 9 /L to 150 × 10 9 /L), and only 3 had severe thrombopenia (platelet count <50 × 10 9 /L). No patients received platelet transfusions, and dual-antiplatelet treatment was withdrawn in patients with severe thrombopenia, except for aspirin when recent stent implantation had been performed. Blood protein decrease was poorly correlated with the severity of platelet count decrease (r 2 = 0.039, p = 0.02). In addition, decreases in hematocrit and blood proteins count averaged 11% and 15%, respectively, while the decrease in platelet count was systematically 24% greater than the plasmatic protein decrease ( Figure 2 ). In addition, platelet change failed to correlate with change in hematocrit (r 2 = 0.007, p = 0.30) and bilirubin (r 2 = 0.004, p = 0.40). Patients with the most severe platelet count decreases (third tertile) had more complex procedures: 5 patients in the third tertile compared with none in the second and first tertiles had prosthesis displacement that required new prosthesis valve implantations in 3 patients (p = 0.04), resulting in lower procedural success in the third tertile. Moreover, patients in the third tertile had more prolonged procedural and x-ray times and larger contrast amounts ( Table 2 ). The differences in procedural duration persisted even after excluding patients with prosthesis displacement (23 ± 13 min in the third tertile vs 17 ± 7 min in the second and first tertiles, p = 0.0008, for x-ray time, and 89 ± 32 vs 71 ± 21 min, respectively, p = 0.0001, for procedure time). Multivariate analysis demonstrated that procedural time (β = 0.23, p = 0.02) and prosthesis migration (β = 0.17, p = 0.03) were the 2 independent predictors of platelet count decrease.
| Variable | Percentage of Platelet Count Decrease by Tertile | p Value | ||
|---|---|---|---|---|
| First | Second | Third | ||
| Age (yrs) | 84 ± 8 | 85 ± 7 | 83 ± 8 | 0.50 |
| Creatinine (μmol/L) | 113 ± 44 | 107 ± 33 | 112 ± 35 | 0.70 |
| Hemoglobin (g/L) | 11.4 ± 1.5 | 12.2 ± 1.5 | 12.1 ± 1.8 | 0.02 |
| Antithrombotic treatment | ||||
| Aspirin | 34 (70%) | 29 (60%) | 26 (54%) | 0.20 |
| Clopidogrel | 35 (71%) | 32 (66%) | 29 (60%) | 0.60 |
| Vitamin K antagonist | 14 (31%) | 18 (37%) | 10 (21%) | 0.20 |
| Procedural characteristics | ||||
| Postdilatation | 6 (13%) | 5 (10%) | 6 (13%) | 0.80 |
| X-ray time (minutes) | 17 ± 7 | 18 ± 7 | 24 ± 13 | 0.002 |
| Procedural time (minutes) | 70 ± 20 | 72 ± 22 | 92 ± 33 | <0.0001 |
| Contrast amount (ml) | 172 ± 77 | 169 ± 77 | 227 ± 128 | 0.006 |
| >1 prosthesis | 0 | 0 | 3 | 0.04 |
| Prosthesis migration | 0 | 0 | 5 | 0.005 |
| Aortic regurgitation ≥2 | 5 (10%) | 10 (21%) | 11 (23%) | 0.30 |
| Biologic characteristics | ||||
| Platelet count at baseline (10 9 /L) | 209 ± 66 | 219 ± 63 | 221 ± 72 | 0.60 |
| Platelet count nadir (10 9 /L) | 172 ± 55 | 146 ± 42 | 108 ± 38 | <0.0001 |
| Platelet count decrease (%) | 18 ± 7 | 33 ± 4 | 50 ± 10 | <0.0001 |
| Protein decrease (%) | 7 ± 15 | 10 ± 8 | 14 ± 9 | 0.01 |
| Hematocrit decrease (%) | 9 ± 19 | 17 ± 19 | 17 ± 10 | 0.02 |
| Bilirubin increase (%) | 21 ± 48 | 44 ± 75 | 31 ± 74 | 0.30 |
| Valve Academic Research Consortium | ||||
| Procedural success | 48 (100%) | 47 (98%) | 42 (88%) | 0.009 |
| 30-day combined safety end point | 4 (8%) | 4 (8%) | 14 (29%) | 0.004 |
| Stroke | 1 (2%) | 4 (8%) | 5 (10%) | 0.20 |
| Major and life-threatening bleeding | 9 (19%) | 14 (29%) | 19 (40%) | 0.08 |
| Major vascular complications | 4 (8%) | 5 (10%) | 10 (21%) | 0.20 |
| Myocardial infarction | 0 (0%) | 2 (4%) | 1 (2%) | 0.40 |
| Other complications | ||||
| In-hospital mortality | 1 (2%) | 3 (6%) | 7 (15%) | 0.06 |
| 30-day mortality | 1 (2%) | 4 (8%) | 7 (15%) | 0.08 |
| MACEs | 10 (21%) | 17 (35%) | 23 (48%) | 0.02 |
| Hemoglobin loss (g/dl) | 1.6 ± 1.1 | 1.9 ± 1.2 | 2.8 ± 1.6 | <0.0001 |
| Red blood cell transfusion | 5 (10%) | 5 (10%) | 17 (35%) | 0.001 |
Kaplan-Meier curves for survival and MACE-free survival are shown in Figure 3 . In-hospital and 30-day mortality were 7% (n = 11) and 8% (n = 12), respectively. Overall, MACEs occurred in 50 patients (35%) during the hospitalization period (11 deaths, 10 strokes and 42 episodes of major bleeding; Table 2 , Figure 4 ). Blood transfusions were delivered in 27 patients, and 8 patients required surgical hemostatic intervention. Bleeding complications, stroke, and in-hospital death were correlated with the severity of platelet count decrease ( Figure 4 ). MACEs were greater in patients with severe platelet count decreases (20% in the first tertile, 38% in the second tertile, and 46% in the third tertile, p = 0.02; Figure 4 ). MACEs were observed in 0 (0%), 35 (39%), and 14 (27%) patients with severe, moderate, and no thrombopenia, respectively. MACEs remained associated with the importance of platelet count decrease even after the exclusion of patients with procedural failure (p = 0.03; Figure 5 ). In addition, the relation between platelet count decrease and MACEs was more significant (p = 0.008) after exclusion of the 30 “learning curve” patients ( Figure 4 ). Finally, blood platelet count decrease was the only predictor of in-hospital MACEs (odds ratio 1.67, 95% confidence interval 1.05 to 2.67, p = 0.03, on multivariate analysis; Table 3 ).
