We read the study by Spitzer et al1 and the editorial by Moreno2 about the usefulness of manual thrombectomy in primary coronary thrombectomy. There is discordance in the literature, with the TAPAS study (Thrombus Aspiration during Percutaneous Coronary Intervention in Acute Myocardial Infarction Study) and 3 meta-analyses3–5 finding it advantageous, but 2 more recent large trials, TASTE (Thrombus Aspiration in ST-Elevation myocardial infarction in Scandinavia) and TOTAL (A Randomized Trial of Routine Aspiration ThrOmbecTomy With PCI Versus PCI ALone in Patients With STEMI Undergoing Primary PCI), suggesting otherwise. Due to its results and sample size, the TASTE trial had a decisive influence on the meta-analysis of 26 trials by Spitzer et al,1 comprising 60.6% of the patients included in the meta-analysis. With the addition of the 10 732 patients from the TOTAL trial, future meta-analyses will take a similar course. However, we believe that there are important aspects that must be taken into account when evaluating the influence of these trials. The TASTE trial was a multicenter, prospective, randomized trial, with patients from the Swedish Coronary Angiography and Angioplasty Registry (SCAAR), and included patients who were candidates for primary percutaneous coronary intervention. Patients gave initial verbal consent and subsequent written consent and were then randomized in a 1:1 ratio to receive primary percutaneous coronary intervention either with or without thrombus aspiration. The primary endpoint was all-cause mortality in the first month and annually thereafter. A total of 11 709 patients attended the centers but, as Moreno pointed out,2 only 7244 (60%) were randomized; the remaining 4580 were included in a parallel registry, 1138 (24.8%) of them undergoing aspiration. The reasons for exclusion were inability to give consent (38%); thrombus aspiration not possible (16%), inappropriate (11%), indicated (7%); and other reasons (28%). The 30-day mortality of the study patients was 2.8% in the thrombus aspiration group vs 3.0% in the control group and was 10.9% vs 10.5% in the 2 registry arms. Finally, although the results were concordant according to Sianos classification of thrombus burden, the study appendix provided the percentages of each study patient group, but not those of the registry. Therefore, we believe that while the TASTE trial has the strengths of being multicenter and prospective, with a very large sample, and based on a positive initiative such as the SCAAR, it also has some limitations. These include not providing the percentage of direct stent implants in each treatment arm, not having a central angiography assessment laboratory, and 2 more aspects that we judge to be of vital relevance: not providing the percentage of types of thrombus for registry patients and, equally important, having a mortality 3 times higher in the parallel registry patients than in the randomized patients.

Similarly, the TOTAL trial included 10 732 patients from 87 hospitals over 4 years of recruitment, with a mean 30.8 cases per year per center. In this study there was no parallel registry of excluded patients. The percentage of patients classified as Killip I was 95.6% in the thrombectomy group and 95.8% in the percutaneous coronary intervention group, and mortality at 180 days was only 3.1% in the thrombectomy group and 3.4% in the percutaneous coronary intervention group without thrombectomy. This information could indicate a significant selection bias.

In our opinion, it is surprising that the analysis of these studies does not mention the low mortality rates.6 Therefore, we believe that the patient profile of both studies bears little resemblance to that seen in everyday practice and consequently the results should be interpreted with caution.