This observational multicenter registry included consecutive patients with symptomatic severe aortic stenosis undergoing TAVI from 8 centers between 2008 and 2021. Patients with baseline thrombocytopenia (< 100*109/L), unavailable platelet counts (pre- or post-TAVI), or intraprocedural death were excluded from the analysis. Eligibility for TAVI and postprocedural management was determined by each Heart Team at the individual center. The decision regarding the type and size of the transcatheter heart valve (THV), access route, and implantation technique was at the discretion of the TAVI operator. Intraprocedural anticoagulation was achieved with unfractionated heparin boluses, starting at 70-100 U/kg at the beginning of the procedure and adjusted to maintain an activated clotting time of >250seconds. Periprocedural antiplatelet therapy was administered according to the protocol in each center. Baseline characteristics, procedural details, clinical outcomes, and follow-up data were prospectively collected in a dedicated centralized database at the coordinating center. Platelet count data were retrospectively collected from laboratory results available in electronic medical records. All patients provided informed consent before the procedure, and the study was conducted in accordance with the institutional review board requirements of the participating centers. This research was undertaken without direct patient involvement; patients were not invited to comment on the study design, develop patient-relevant outcomes, or interpret the results. There was no dedicated funding for this study.

Postprocedural thrombocytopenia was defined as a ≥ 50% decrease in platelet count compared with baseline within 0 to 10 days after TAVI.6 The percentage decrease in platelet count (DPC) was calculated as: [(baseline platelet count - nadir platelet count)/baseline platelet count]*100]. Early platelet count nadir was defined as the lowest platelet count within 3 days post-TAVI, and late nadir was defined as occurring from day 4 onward.4 Clinical outcomes were defined according to the Valve Academic Research Consortium (VARC-2) criteria. To determine the presence of prosthesis patient mismatch, previously defined predicted effective orifice area for each valve type and size were used, and indexed by body surface area.7,8 The primary endpoint was 30-day all-cause mortality, and the secondary endpoints were procedural safety and 2-year all-cause mortality.

The data are reported as mean and standard deviation for continuous variables and as numbers and percentages (%) for categorical variables. All patients were classified according to the presence or absence of postprocedural thrombocytopenia, and comparison were performed using the 2-sided Student t test for continuous variables and the chi-square test for categorical variables, as appropriate.

A subanalysis was conducted to assess the impact of platelet count dynamics. Patients were classified into 3 groups: group 1, without thrombocytopenia; group 2, thrombocytopenia with early nadir; and group 3, thrombocytopenia with late nadir. To avoid selection bias, only patients with a hospital stay of ≥ 4 days (and with available hemogram beyond day 3) and those with an increase in platelet count after the post-TAVI nadir were included in this subanalysis.

Survival analysis was performed using a Kaplan-Meier survival function and the curves were compared using the log-rank test. Logistic regression was used to assess predictors of thrombocytopenia and 30-day mortality. The variables associated with these outcomes in the univariable analysis (with a P value of <.100) were included in the multivariable logistic model. In addition, a propensity score matching analysis was performed to obtain 2 matched groups with and without thrombocytopenia, adjusted for baseline and procedural characteristics. The primary and secondary endpoints were analyzed in this matched cohort to assess the clinical impact of thrombocytopenia. Details of the propensity score matching method are provided in the supplementary data.

A subgroup analysis was performed to evaluate the impact of thrombocytopenia on primary and secondary endpoints in subgroups defined by sex (male and female), access type (transfemoral and nontransfemoral), different THV types, and TAVI procedures performed before and after 2015. The year 2015 was chosen as it represented the mid-point of the inclusion period. Due to advances in technique and data from large, randomized studies, patients from 2015 onwards represent a population more comparable to current clinical practice in TAVI, which was the rationale for this subgroup analysis.

P values of less than .05 were considered statistically significant. All data were analyzed using Stata 14 (StataCorp, College Station, United States).

Laboratory findings are presented in table 2. The mean baseline platelet count was 209±70 *109/L and the mean post-TAVI nadir was 136±54 *109/L, occurring at 2.5±1.5 days postprocedure. The percentage DPC was 34.3%±15.4. Postprocedural thrombocytopenia (a decrease of ≥ 50% from baseline) occurred in 578 (14.8%) patients. A subsequent increase in platelet count was observed in 2461 (63.0%) patients after 4.2±1.8 days post-TAVI. The increase in platelet count was observed in 78.8% of patients with a hospital stay of ≥ 4 days. The mean platelet count at discharge was 171±73 *109/L, and the percentage decrease in platelet count at discharge relative to baseline was 16.4%±29.0. Only 19.6% of patients reached a platelet count equal to or higher than baseline values.

A total of 3041 patients met the criteria of ≥ 4 days hospital stay (n=2853) or a postnadir increase in platelet count within hospital stay <4 days (n=188). In this population, early thrombocytopenia occurred in 323 (62.7%) patients, and late nadir occurred in 192 (37.3%). Platelet count dynamics are presented in figure 1 and figure 2. Platelet count at 3 months postprocedure was obtained in 2049 (54.9%) patients, with a mean platelet count of 200±73 *109/L. Thrombocytopenia was observed in only 41 (2%) patients at this follow-up. Among patients with postprocedural thrombocytopenia, only 14 (4.7%) showed persistent thrombocytopenia at follow-up.

Figure 1.

Dynamics in platelet count after transcatheter aortic valve implantation. Platelet count change within 7 days post-TAVI overall (A) and according to early and late nadir (B). TAVI, transcatheter aortic valve implantation; TP, thrombocytopenia.

(0.22MB).

Figure 2.

Central illustration. Platelet count pattern following TAVI and clinical impact of acquired thrombocytopenia. A: dynamics of platelet count within 7 days post-TAVI overall and according to early and late nadir. B: impact of thrombocytopenia on 30-day mortality by subgroups. C: 2-year all-cause mortality according to early and late thrombocytopenia. TF: transfemoral. THV: transcatheter heart valve. TP, thrombocytopenia.

(0.56MB).

Table 3 show periprocedural complications and clinical outcomes related to the occurrence of thrombocytopenia. Overall 30-day mortality was higher in patients with thrombocytopenia (2.0 vs 8.5%; odds ratio [OR] 4.6; 95% confidence interval [95%CI], 3.1-6.7; P <.001). Other variables associated with 30-day mortality were age, STS score, major vascular complications, major and life-threatening bleeding, acute kidney injury, periprocedural stroke, and severe aortic regurgitation. The adjusted OR for 30-day mortality associated with thrombocytopenia was 2.3 (95%CI, 1.3-4.2; P=.004, table 1 of the supplementary data). Early safety and device success at 30 days were significantly higher in the group without thrombocytopenia (77% vs 52.1%; P <.001, and 89.5% vs 75%; P <.001, respectively).

Both early and late thrombocytopenia were associated with worse 30-day clinical outcomes (table 2 of the supplementary data) and increased 30-day mortality compared with patients without thrombocytopenia (adjusted OR, 2.1; 95%CI, 1.0-4.2; P=.041, and 5.4; 95%CI, 2.9-10.0; P <.001), respectively).

Follow-up was obtained in 3775 (96.5%) patients and the mean follow-up time was 28.5±25.5 months. The 2-year all-cause mortality was 18.9% (30.2% vs 16.8% in patients with and without thrombocytopenia, respectively: hazard ratio [HR], 2.2; 95%IC, 1.8-2.7; P <.001), as shown in figure 3A. After a 30-day landmark analysis this difference remained significant (23.8% vs 15.2%; HR, 1.8; 95%CI, 1.4-2.2; P <.001), as shown in figure 3B. The 2-year cardiovascular mortality was also higher in patients with thrombocytopenia before and after the 30-day landmark analysis (figure 3E,F).

Figure 3.

Two-year all-cause mortality and cardiovascular mortality. Kaplan-Meier graph of 2-year all-cause mortality and after 30-day landmark analysis according to thrombocytopenia (A, B) and early and late nadir (C, D). Two-year cardiovascular mortality and after 30-day landmark analysis according to thrombocytopenia and early and late nadir (E-H). CV, cardiovascular; TP, thrombocytopenia.

(0.79MB).

Among patients with thrombocytopenia, 2-year mortality was significantly higher in those with late nadir compared with early nadir (38.6% vs 23.8%, p<.001, figure 3C and figure 2 – central illustration). This difference remained higher after the 30-day landmark analysis (20.4% vs 30.1%; P=.008, figure 3D). Compared with early nadir, late nadir was associated with increased 2-year cardiovascular death (figure 3G), with the difference mainly observed within the first 30 days (figure 3H).

Factors independently associated with thrombocytopenia are presented in table 4. Baseline characteristics that increased the risk of thrombocytopenia were body surface area, absence of diabetes, peripheral vascular disease, and lower estimated glomerular filtration rate (eGFR). Other factors independently associated with thrombocytopenia were procedural aspects, such as TAVI performed in the first half of the study (< 2015), nontransfemoral access, use of the Portico (Abbott, United States) THV, and adverse procedural outcomes, such as major vascular complications.

Regarding the type of THV, no significant differences were observed between self-expanding (SEV) and balloon-expandable valves (BEV). However, patients receiving the CoreValve (Medtronic, United States) or Acurate Neo (Boston Scientific, United States) THV had the lowest rates of thrombocytopenia (8.5% and 9.7%, respectively), whereas recipients of the SAPIEN (Edwards System, United States) valve had an intermediate rate (14.2%), and those receiving the Portico valve had the highest occurrence of thrombocytopenia (33.8%) (figure 1 of the supplementary data). Both early and late thrombocytopenia were significantly higher in patients with the Portico valve compared with other THVs (early: 21.3% vs 9.7%; P <.001, and late: 18.3% vs 5.3%, P <.001).

Regarding clinical outcomes, thrombocytopenia did not significantly impact 30-day or 2-year all-cause and cardiovascular mortality among patients receiving the Portico valve (figure 4).

Figure 4.

Impact of post-TAVI thrombocytopenia on clinical outcomes by subgroup analysis. Forest plot representing the clinical impact of thrombocytopenia on procedural safety (A), 30-day mortality (B), and 2-year all-cause (C) and cardiovascular mortality (D) in different subgroups. TAVI, transcatheter aortic valve implantation; TF, transfemoral, THV; transcatheter heart valve, TP; thrombocytopenia.

(0.3MB).

Late thrombocytopenia was associated with increased surgical risk, as assessed by the STS score, and with postprocedural complications including major or life-threatening bleeding, moderate-to-severe aortic regurgitation, and stage 2 or 3 acute kidney injury (AKI) (table 3 of the supplementary data).

After the propensity score matching analysis, 2 matched groups of 543 patients each, with and without thrombocytopenia were obtained. Baseline and procedural characteristics were balanced between the groups (figure 2 of the supplementary data and table 4 of the supplementary data). The 30-day all-cause mortality and procedural safety remained statistically significantly different between the matched groups with and without thrombocytopenia [39 (7.3%) and 13 (2.5%), P <.001, and 289 (53.6%) and 379 (71.4%), P <.001, respectively, table 5 of the supplementary data].

Moreover, 2-year all-cause and cardiovascular mortality were significantly higher in the propensity score matched group with thrombocytopenia compared with those without thrombocytopenia (29.2% vs 16.7%; HR, 2.0; 95%CI, 1.5 - 2.7; P <.001, and 16.5% vs 7.2%, HR, 2.7; 95%CI, 1.7-4.4; P <.001, respectively, table 5 of the supplementary data and figure 3 of the supplementary data).

The clinical impact of thrombocytopenia on procedural safety, 30-day mortality, and 2-year all-cause and cardiovascular mortality was consistently observed in all subgroups according to sex, year of TAVI, and access route. However, in the Portico THV subgroup no significant impact of thrombocytopenia on short or mid-term mortality was observed (figure 4).

We acknowledge several limitations of this study, which are partly due to its retrospective design. First, no additional complementary testing was conducted to clarify the etiology of thrombocytopenia, particularly in cases of late thrombocytopenia or heparin-induced thrombocytopenia. Second, despite multivariable and propensity score matching analyses, confounding factors related to the long inclusion period, procedural approaches, and peri-procedural complications cannot be entirely ruled out. Third, in-hospital infections and sepsis were not prospectively collected and, therefore, could not be analyzed as potential contributors to acquired thrombocytopenia. Fourth, shorter hospital stays for patients undergoing minimally invasive TAVI may have led to bias due to the absence of platelet count measurements postdischarge. Finally, the multicenter design and extended inclusion period introduced variability in technical approaches and pharmacological therapies, including antithrombotic regimens. However, this also enhances the external validity of our results, as participating centers adhered to the guidelines of the American and European Cardiology Societies.