Blood Transfusions and Low Platelets Flag Bleeding Risk in Prostate Cancer Patients on Antithrombotics

For men with advanced prostate cancer approaching the end of life, a routine decision about whether to continue blood thinners may carry more weight than clinicians have traditionally assumed. A new retrospective cohort study from a specialized palliative care unit in Stockholm, Sweden, published in Cancer Reports, has identified specific clinical markers that flag which patients are most vulnerable to serious bleeding while taking antithrombotic medications during their final year of life. The findings arrive at a time when palliative care teams worldwide are increasingly questioning whether medications designed to prevent clots still make sense for patients whose remaining lifespan is measured in months.

Antithrombotic treatments, which include antiplatelet drugs such as aspirin and clopidogrel as well as anticoagulants like low-molecular-weight heparin, warfarin, and direct oral anticoagulants, are remarkably common in specialized palliative cancer care. Clinical guidelines suggest that most patients with cancer who develop thrombosis benefit from long-term antithrombotic treatment, regardless of adverse effects. Yet the research team behind the new study had previously shown, in a retrospective analysis of palliative care patients, that the risk of bleeding in patients using antithrombotics during the last year of life was high, independent of age, indication, and type of antithrombotic, while the risk of thrombosis after deprescribing was very low. Intriguingly, that earlier work also revealed that patients with advanced prostate cancer had a significantly increased risk of bleeding compared to patients with other cancer types and non-cancer diagnoses. The primary aim of the present study was therefore to dig deeper into what makes men with prostate cancer on antithrombotics especially prone to bleeding.

The study was designed as a retrospective single-center cohort investigation, in which the researchers reviewed medical records for all deceased patients who had been admitted to ASIH Stockholm Södra, a Stockholm palliative care unit providing both specialized home care and in-patient specialized palliative care, from January 1, 2016, until December 31, 2018. The original cohort had screened all admitted patients for use of antithrombotics, classified under ATC code B01, during the last year of life, excluding patients with prophylactic antithrombotic treatment after surgery. For this sub-study, the researchers extracted all patients with prostate cancer, both with and without antithrombotic treatment, from the original cohort. The study was approved by the Swedish Ethical Review Board.

The resulting cohort consisted of 100 men with prostate cancer, of whom 64, or 64 percent, were treated with antithrombotics during the last year of life. The mean age at death was 79.1 years. The most common antithrombotic treatment was low-molecular-weight heparin, used by 31 patients, or 48.4 percent, followed by antiplatelet drugs in 27 patients, or 42.2 percent. Only 6 patients, or 9.4 percent, had other oral drugs such as warfarin or direct oral anticoagulants. Prophylaxis and atrial fibrillation were the most common indications for antithrombotic treatment, with prophylaxis defined as cases where no clear indication could be found in the medical records.

Among those treated with antithrombotics, 24 patients, or 37.5 percent, experienced clinically significant bleeding during their final year. The most common source of bleeding was hematuria, blood in the urine, which affected 12 patients, followed by nasal bleeding and bleeding from the skin, each affecting four patients. One patient suffered a central nervous system bleed. In the entire cohort, 36 men with prostate cancer had no treatment with antithrombotics, and eight of these men, or 22.2 percent, experienced some form of bleeding, most commonly hematuria followed by nasal bleeding. Perhaps surprisingly, there was no statistically significant difference in the risk of bleeding between those treated with antithrombotics and those not treated, with a relative risk of 0.59 and a 95 percent confidence interval of 0.29 to 1.23, and a p-value of 0.12. The researchers caution that this might be due to lack of power in the limited cohort, but their data do not support a clear increase in bleeding risk attributable to antithrombotics alone.

To identify clinical markers of vulnerability, the researchers used log-binomial generalized linear models to calculate relative risks and 95 percent confidence intervals for each variable potentially related to bleeding, including age, indication for antithrombotics, and treatment with steroids or non-steroidal anti-inflammatory drugs. Factors associated with increased risk of bleeding were the need for blood transfusion during the last year of life, with a relative risk of 3.15 and a 95 percent confidence interval of 1.34 to 7.39, and low thrombocyte count, defined as fewer than 50 times 10 to the ninth per liter, with a relative risk of 2.41 and a 95 percent confidence interval of 1.39 to 4.16. In addition, patients receiving antithrombotics for deep vein thrombosis treatment had a higher risk of bleeding compared to those on prophylactic treatment, with a relative risk of 2.44 and a 95 percent confidence interval of 1.21 to 4.94. Notably, age, type of antithrombotic, and use of steroids or NSAIDs were not associated with increased bleeding risk.

The relationship between blood transfusion and bleeding is probably one of reverse causality, the researchers explain, since bleeding or bone marrow failure might occur before the transfusions. Nevertheless, transfusion need might serve as a practical marker for when antithrombotic treatment should be evaluated. Earlier studies show that high rates of significant bone marrow failure, including anemia and thrombocytopenia, in treatment-refractory patients with prostate cancer predict an unfavorable prognosis, which is in concordance with the new results showing that low thrombocyte count was associated with a high risk of bleeding. Thus, when thrombocytopenia occurs, it might be time to evaluate whether antithrombotic treatment should continue. Hematuria, which is well established as common in prostate cancer, was also the most common source of bleeding in this study, and could be one explanation for the high bleeding risk in men with prostate cancer compared to patients with other diagnoses. Previous work by Ajabnoor and colleagues has shown that men with prostate cancer and atrial fibrillation generally have an increased risk for bleeding, in line with the team’s earlier findings.

The study also examined the effects and timing of deprescribing antithrombotics, with striking results. No patients developed any cardiovascular complications, such as myocardial infarction, stroke, or systemic embolism, after deprescribing antithrombotics during the last year of life. For 32 patients, or 50 percent, antithrombotics continued until the last days of life and were deprescribed zero to six days before death. Antithrombotics were not deprescribed at all for 14 patients, or 22 percent. For those who continued treatment until the very last days, approximately 43 percent received it orally, though it is not known whether they were able to swallow the medication, as most patients cannot take oral medications during the last days of life.

The incidence of bleeding in the present study, at 37.5 percent of all men with prostate cancer treated with antithrombotics, is higher than in earlier studies, including the team’s original study where the incidence was 16.1 percent in all patients treated with antithrombotics, and a French study by Tardy and colleagues where the incidence of clinically relevant bleeding was 9.8 percent. However, those studies included palliative care patients with all kinds of diagnoses, not only men with prostate cancer, which may partly explain the difference. To their knowledge, the researchers note, this is the first study to evaluate clinical markers for vulnerability to bleeding in men with prostate cancer and treatment with antithrombotics in palliative care.

The study has several limitations that warrant careful interpretation. First, the sample size was small, which may affect the results in terms of power, and the effect estimates should be interpreted with caution, since the confidence intervals are wide and there is high potential for both type I and type II errors. Second, the study is based on retrospective data, and confounding factors and reverse causality might be present, so no causal relationships can be drawn. Due to the small sample size and retrospective, non-interventional design, no firm conclusions about safety after deprescribing can be drawn, which needs to be evaluated in a future prospective study. Moreover, laboratory values were obtained up to six months before death and may not reflect the status at the time of bleeding, a potential source of misclassification. Hematuria is also common in patients with prostate cancer regardless of whether they are treated with antithrombotics or not. Finally, the results are based on data from a single Swedish specialized palliative care unit, so they may not be generalized to other healthcare systems or less advanced care settings. Still, the researchers conclude that given the increased risk of bleeding and the often limited life expectancy of these patients, the potential benefits of preventing thromboembolic events must be weighed against the potential harms of antithrombotic treatment in each case, with clinical markers such as low thrombocyte count and need for transfusions offering clinicians concrete signposts for when that evaluation should begin.

Subject of Research: Clinical Markers of Vulnerability for Bleeding in Men With Prostate Cancer and Antithrombotic Treatment

Article Title: Clinical Markers of Vulnerability for Bleeding in Men With Prostate Cancer and Antithrombotic Treatment

Article References: Björkhem‐Bergman, L., Hedman, C., Szilcz, M., & Frisk, G. (2026). Clinical Markers of Vulnerability for Bleeding in Men With Prostate Cancer and Antithrombotic Treatment. Cancer Reports, 9(10), Article e70720. https://doi.org/10.1002/cnr2.70720

Image Credits: AI Generated

DOI: 10.1002/cnr2.70720

Keywords: palliative care, prostate cancer, antithrombotics, bleeding risk, thrombocytopenia, deprescribing, end-of-life care, Cancer Reports, Clinical, Markers, Vulnerability, Bleeding