Author + information
- Marc Schindewolf, MD∗ (, )
- Norbert Banik, PhD and
- Edelgard Lindhoff-Last, MD
- ↵∗Swiss Cardiovascular Center, Division of Vascular Medicine, University Hospital Bern, Freiburgstrasse 4, Building PKT1 C 421, 3010 Bern, Switzerland
Our retrospective registry was planned to describe treatment effects within different treatment strata and not as a comparison between them (1). Therefore, data must be interpreted with caution in regard to causality and outcome prediction by differences in baseline characteristics. The definition of universal predictive risk factors for direct matching would result in sample sizes per group that are too small to analyze. Alternative approaches, such as propensity score adjustments that would circumvent the technical problem of too-small sample sizes, would have only shifted our study aim from its descriptive focus to a more strictly comparative character. Such an undue focus is exactly what we wanted to avoid.
Nevertheless, an analysis of differences in baseline characteristics (Table 1) shows no statistical difference in the frequencies of positive laboratory heparin-Induced thrombocytopenia (HIT) diagnostics between the different anticoagulants. Thus, in regard to the aspect of the possibility of underlying HIT, the different therapy strata were comparable. Other significant baseline characteristics, for example, hypertension, venous insufficiency are rather unlikely to have a relevant influence on the assignment to one or another therapy group or even therapy outcome. Significant differences might also be due to false-positive findings when multiple subgroup comparisons are performed (the “multiplicity problem”) (2). This also applies to significant baseline characteristics that might influence thrombosis/bleeding risk and therefore therapy outcome, for example, cancer, or infection.
HIT incidence is influenced by various factors (e.g., patient population, heparin formulation, and dose) (3) and have been reported previously regarding our study population (4). However, these factors do not affect the validity of our results, because patients were enrolled after HIT had already clinically been suspected. For example, heparin treatment preceding the HIT episode in fondaparinux-anticoagulated versus alternatively anticoagulated patients had only an influence on the baseline risk to develop HIT, but not after robust HIT suspicion had already been established and heparin stopped. Furthermore, fondaparinux-treated patients had on average a higher 4Ts score compared with non-fondaparinux therapy groups (6 points [high risk] vs. 5 points [intermediate risk]).
Many study patients did not have “true” HIT based on post hoc analysis of laboratory results (5). However, initially, even these patients did have a strong clinical HIT suspicion (4Ts score ≥4 points) and therefore, an indication for alternative anticoagulation before specific HIT diagnostics is available (3). Fondaparinux seems effective and safe in these patients. Another criticism of Dr. Zhu and colleagues was that outcome in fondaparinux-treated patients might be favorable due to the low likelihood of “true” HIT, which is corroborated by double-negative HIT diagnostics. In order to overcome this drawback of combining patients with different HIT likelihoods, we have identified 35 patients with “true” HIT by meticulously matching clinical parameters with laboratory HIT diagnostics. Also, in this “true” HIT subgroup, thrombotic and bleeding risk during fondaparinux treatment was low.
Please note: The study was supported by GlaxoSmithKline. The contract research organization and the physicians who were screening the medical records obtained financial compensation for data acquisition and data management from GlaxoSmithKline. The sponsor had no influence on the study design; on the collection, analysis, and interpretation of data; in the writing of the manuscript, and in the decision to submit the manuscript for publication. There were no agreements concerning confidentiality of the data between the sponsor and the authors or the institutions named in the credit lines; all collected data were fully disclosed. Dr. Schindewolf has received speaker fees from Abbott, Aspen, Boston Scientific, Bristol-Myers Squibb, Daiichi-Sankyo, GlaxoSmithKline, and Sanofi; has received travel grants from Bard, Bayer Healthcare, Boston Scientific, Bristol-Myers Squibb, Medtronic, and Terumo; has received research grants from Cook, Daiichi-Sankyo, and Terumo; and is a member of the advisory boards of and a consultant for Bayer Healthcare, Bristol-Myers Squibb, Daiichi-Sankyo, and Sanofi. Dr. Banik is a past employee of GlaxoSmithKline. Dr. Lindhoff-Last has received speaker fees from Bayer Healthcare, Boehringer Ingelheim, GlaxoSmithKline, and Sanofi; and is a member of the advisory boards of Bayer Healthcare, Boehringer Ingelheim, and GlaxoSmithKline.
- 2018 American College of Cardiology Foundation
- Schindewolf M.,
- Steindl J.,
- Beyer-Westendorf J.,
- et al.
- Linkins L.A.,
- Dans A.L.,
- Moores L.K.,
- et al.
- Warkentin T.E.