Meta-Analysis Will Compare Viscoelastic Guidance With Conventional Coagulation Testing in ICU Bleeding
Bleeding remains a major driver of morbidity and mortality in intensive care, but the way clinicians measure and respond to coagulopathy still varies widely across hospitals. This BMJ Open protocol sets out a systematic review and meta-analysis of viscoelastic haemostatic assay guided transfusion versus standard laboratory coagulation testing in critically ill adults who are not in a cardiac surgery setting and who have active bleeding or clinically significant coagulopathy.
The question is practical at the bedside. In the unstable ICU patient, transfusion decisions must often be made before conventional coagulation tests can fully capture the dynamics of clot formation, clot strength, and fibrinolysis. Viscoelastic assays such as thromboelastography and rotational thromboelastometry offer a more immediate and global view of haemostasis, which may allow clinicians to target blood products more precisely rather than treating numbers in isolation.
Why Conventional Coagulation Panels May Miss the Bleeding Phenotype
Standard tests such as prothrombin time, activated partial thromboplastin time, and fibrinogen concentration provide useful information, but they are static measures. They do not fully reflect the complex interaction among platelets, fibrin formation, clot firmness, and fibrinolysis that drives real-time bleeding in critical illness. In a patient who is actively losing blood, a delay in targeted correction can mean ongoing transfusion, worsening shock, and a higher risk of organ dysfunction.
Viscoelastic testing is attractive because it can compress the diagnostic window. A single bedside assay may guide the choice between plasma, platelets, fibrinogen replacement, or other haemostatic products more directly than conventional laboratory testing. The biological logic is straightforward: if clot initiation, propagation, and stability can be measured in real time, transfusion can be matched more closely to the defect actually present.
Review Design Targets a General ICU Population, Not a Specialised Haemostatic Niche
This protocol deliberately separates general non-cardiac critical care from more specialised haemostatic environments. That design choice matters because bleeding after cardiac surgery, trauma, liver transplantation, and other highly protocolised settings can follow different transfusion pathways and produce different evidence signals. Restricting eligibility to critically ill non-cardiac adults with a clinically relevant treatment decision should make the review more applicable to mainstream intensive care practice.
The planned search is broad and methodical. It combines controlled vocabulary with free-text terms, applies no language, date, or study-design filters, and includes known-item validation of the PubMed strategy against sentinel studies. A complete reproducible search strategy is provided in the supplementary material, which should help readers judge whether the final evidence base was assembled transparently and without unnecessary restriction.
Mixed Study Designs Could Expand the Evidence Base and Complicate Interpretation
Both randomised and non-randomised comparative studies will be eligible. That will widen the net in a field where the number of direct head-to-head comparisons may be small, but it also introduces a familiar methodological tension. Randomised trials can better balance measured and unmeasured confounding, while observational studies may reflect everyday practice more closely yet remain vulnerable to treatment-selection bias, severity imbalance, and protocol drift.
The review plans to use Risk of Bias 2 for randomised studies and ROBINS-I for non-randomised interventions. Those tools are appropriate for a question like this because transfusion practice is highly contingent on illness severity, bleeding source, local algorithms, and clinician preference. In real ICU care, patients are not assigned to viscoelastic testing by chance alone; they are often triaged by circumstance, which can distort apparent treatment effects if the underlying case mix is not carefully handled.
A small evidence base could leave the final synthesis with wide confidence intervals. If the available studies diverge in assay type, transfusion thresholds, or co-interventions, pooled estimates may be unstable or impossible to combine meaningfully.
Heterogeneity May Be the Central Limiting Feature
The protocol explicitly anticipates clinical heterogeneity, variation in VHA algorithms, and confounding in observational studies. Those issues are not peripheral. They are likely to determine whether the meta-analysis can answer a clean clinical question or only map the boundaries of the current literature.
Viscoelastic pathways are rarely uniform across institutions. Some centres may use specific trigger values for fibrinogen or platelets, while others may integrate assay results into broader massive bleeding or transfusion bundles. Even when two hospitals both claim to use VHA-guided transfusion, the actual decision tree may differ enough to dilute a treatment effect or create the illusion of one where none exists.
One sentence captures the practical challenge: a faster assay does not automatically produce a better transfusion strategy.
What Clinicians Will Want From the Final Synthesis
The main clinical value of the completed review will not depend solely on whether viscoelastic-guided transfusion reduces mortality. In critical care, a meaningful benefit could also emerge through lower blood product exposure, fewer unnecessary transfusions, improved haemostatic control, or reduced downstream complications from overtransfusion. The protocol is designed to capture comparative evidence across these endpoints if the literature allows it.
Still, the review will need to distinguish signal from convenience. VHA-based management may look superior in settings where conventional testing is delayed or incomplete, yet the same approach may offer less advantage where laboratory turnaround is rapid and transfusion decisions are already tightly protocolised. A bedside test is only as useful as the treatment algorithm it drives.
Evidence Gaps Are Already Visible Before the Search Is Complete
The protocol makes clear that the current literature may not be robust enough for confident quantitative pooling across all domains of interest. That is a realistic admission rather than a weakness of the project. It reflects the structure of the field itself: a narrow comparative evidence base, changing assay platforms, and significant variation in how clinicians translate test patterns into blood product therapy.
If the final review identifies consistent benefit across transfusion use, bleeding control, or patient-centred outcomes, it could strengthen the case for broader adoption of viscoelastic testing in non-cardiac intensive care. If the evidence remains fragmented, the result will be equally useful, because it will define where the uncertainty persists and where future trials need to be more uniform in design.
For now, the protocol establishes the framework. The unresolved issue is whether the available comparative studies are sufficiently aligned in population, assay strategy, and outcome reporting to support a credible estimate of effect rather than a collage of incompatible ICU practices.