Desmopressin Fails to Help Brain Bleed Patients on Antiplatelet Drugs, Bayesian Analysis Warns of Possible Harm

For millions of people living with heart disease, a daily aspirin or clopidogrel tablet is a lifeline against heart attacks and strokes. But the same drugs that keep platelets from clumping together can turn a brain hemorrhage into a catastrophe. When a blood vessel bursts spontaneously in the brain of a patient on antiplatelet therapy, the blood refuses to clot properly, the hematoma keeps growing, and mortality climbs. For years, clinicians have reached for desmopressin, a synthetic hormone analogue that promises to supercharge platelet function at the bleeding site. Now, a comprehensive new synthesis of the evidence suggests that this widely considered rescue drug may not help at all, and may even cause harm.

The study, published in the journal Neurocritical Care, is a systematic review and Bayesian meta-analysis led by Paweł Łajczak of the Medical University of Silesia together with collaborators in Brazil and Poland. The team searched five major databases from their inception to January 2026 and, after rigorous screening, included seven comparative studies encompassing 731 patients with antiplatelet therapy-associated intracerebral hemorrhage, of whom 281 received desmopressin. Their conclusion is stark: there is currently no convincing evidence that desmopressin improves survival or functional recovery in these patients, and several statistical signals raise genuine concern for harm.

Desmopressin, also known as DDAVP, is a selective agonist of vasopressin V2 receptors. When administered, it triggers the release of von Willebrand factor and Factor VIII from endothelial stores, both of which are critical for platelets to adhere to damaged vessel walls and form a stable plug. Early exploratory work, including a 2014 study showing improved platelet activity after desmopressin administration in intracerebral hemorrhage, made the drug an attractive candidate for limiting hematoma expansion. The logic seemed airtight: if antiplatelet drugs paralyze platelets, and desmopressin re-arms them, bleeding should slow and outcomes should improve. The new analysis suggests the biology of spontaneous brain hemorrhage is far less cooperative than that simple model implies.

The stakes are considerable. Patients on dual antiplatelet therapy face a more than 40 percent higher risk of intracerebral hemorrhage compared with those taking aspirin alone, and studies show that patients who bleed while on antiplatelet agents have larger hemorrhage volumes at 24 hours and higher mortality than those who are not. Unlike anticoagulant reversal, the antiplatelet effect cannot be undone immediately, which is precisely why a fast-acting prohemostatic agent like desmopressin has been so appealing in the emergency setting.

Methodologically, the new review stands out for its statistical approach. Because only a small number of studies and patients exist in this literature, traditional frequentist meta-analysis risks overstated conclusions and unstable estimates. The authors instead employed a Bayesian random-effects model, run in the bayesmeta package in R, using three different priors: a vague prior for the overall effect size, informative log-normal priors derived from validated predictive distributions published by Turner and colleagues, and sensitivity analyses with weakly informative and vague half-normal priors for heterogeneity. Rather than declaring results simply significant or nonsignificant, the Bayesian framework yields posterior probability distributions, which quantify how likely an effect is to point in a given direction given the data and prior knowledge. The study was prospectively registered in PROSPERO and followed Cochrane Handbook and PRISMA guidelines.

The headline results were uniformly negative for benefit. Desmopressin showed no reduction in overall mortality, with an odds ratio of 1.23 and a 95 percent credible interval of 0.79 to 1.91, and no reduction in in-hospital mortality, with an odds ratio of 1.41 and a credible interval of 0.88 to 2.26. Good functional outcome, defined as a modified Rankin Scale score of 0 to 3 at last follow-up, was actually numerically worse in the desmopressin group, with an odds ratio of 0.66 and a credible interval of 0.34 to 1.31. Hematoma expansion over 33 percent, the mechanistic endpoint that desmopressin was supposed to prevent, showed a nonsignificant trend toward benefit with an odds ratio of 0.89, far short of credible statistical evidence. Adverse events, thromboembolic complications, and hyponatremia all trended toward higher rates with the drug.

The most unsettling findings emerged from the posterior distribution analyses. The model estimated a 74.88 percent probability that desmopressin increases thromboembolic events, roughly a 70 percent probability of increased adverse events overall, and an 82.28 percent likelihood that the drug is associated with lower odds of achieving a good functional outcome. The authors are careful to note that these probabilities should not be read as proof of harm, since the 95 percent credible intervals cross the null-effect line and no prespecified interpretive thresholds exist. But in the absence of any demonstrated baseline benefit, uncertain signals of potential harm carry considerable weight. The thrombotic concern has a plausible mechanism: patients who need antiplatelet therapy already carry high vascular morbidity, and flooding their circulation with von Willebrand factor and Factor VIII could tip the balance toward clot formation in coronary or cerebral vessels.

Only one of the seven included studies was randomized: the UK-based DASH trial, a phase 2, placebo-controlled feasibility study published in The Lancet Neurology in 2023. DASH found no significant differences in hematoma expansion, 21 percent with desmopressin versus 23 percent with placebo, nor in mortality or neurological outcomes. Intriguingly, its safety profile diverged sharply from the observational literature, with only one thromboembolic event in each arm, roughly 4 percent, whereas retrospective studies had reported much higher thrombotic rates. The authors suggest this discrepancy highlights how heavily confounded retrospective data can be, and they argue that fully powered phase 3 trials are essential before any definitive verdict on safety can be reached. The DASH trial itself was judged at low risk of bias, while the six nonrandomized studies showed concerns primarily around confounding, participant selection, and missing data, and all primary outcomes in the review were graded as low certainty under the GRADE framework.

Why does desmopressin fail to deliver in this population when it has shown promise in traumatic brain injury? The authors point to the pathophysiology of spontaneous intracerebral hemorrhage, where high blood pressure in small vessels physically disrupts platelet clot formation in ways that extra von Willebrand factor cannot fully rescue. Timing is another likely culprit: because hemorrhage strikes suddenly, many patients reach the intensive care unit too late for an ultra-early dose, and the optimal administration window may simply be too narrow to exploit. The review also underscores that desmopressin is only one variable among many. In one included cohort, the time from admission to first CT scan, platelet transfusion, and age were significant predictors of hematoma expansion, while in another multivariate model only increased age and the presence of intraventricular hemorrhage independently predicted poor short-term neurological outcome, with the intervention itself failing to reach significance.

The authors argue that the field must stop treating antiplatelet-associated hemorrhage as a homogeneous entity. Patients take different antiplatelet agents with different degrees of platelet inhibition and carry wildly different comorbidity profiles, and it remains unknown whether specific subpopulations might genuinely benefit from desmopressin. Future trials, they contend, should compare desmopressin directly against other prohemostatic agents and focus on identifying responsive subsets of patients tied to specific antiplatelet regimens. For now, the practical message for clinicians is sobering: routine desmopressin administration in spontaneous brain hemorrhage on antiplatelet therapy lacks demonstrated clinical benefit, and the possibility of thromboembolic harm demands that the drug be studied, not assumed, before it becomes standard of care.

Subject of Research: Effectiveness and safety of desmopressin in spontaneous intracerebral hemorrhage among patients on antiplatelet therapy

Article Title: Use of Desmopressin in Spontaneous Nontraumatic Intracerebral Hemorrhage Among Patients Taking Antiplatelet Therapy: Systematic Review and Bayesian Meta-analysis

Article References: Use of Desmopressin in Spontaneous Nontraumatic Intracerebral Hemorrhage Among Patients Taking Antiplatelet Therapy: Systematic Review and Bayesian Meta-analysis. (n.d.). https://doi.org/10.1007/s12028-026-02672-3

Image Credits: AI Generated

DOI: 10.1007/s12028-026-02672-3

Keywords: desmopressin, DDAVP, intracerebral hemorrhage, antiplatelet therapy, Bayesian meta-analysis, hematoma expansion, thromboembolic events, stroke, neurocritical care, von Willebrand factor, systematic review, DASH trial