A-V shunt thrombosis in NHPs: interpreting flow, thrombus and bleeding endpoints
The amount of clot recovered at the end of an arteriovenous (AV) shunt thrombosis study is one readout among several. In a non-human primate (NHP) AV shunt model, blood runs through a defined extracorporeal test segment under flow. The setup supports assessment of flow behavior and thrombus formation on a blood-contacting surface, and complementary hemostasis endpoints describe the safety side of an antithrombotic response.
These measurements answer different questions. Flow indicates whether the developing thrombus is affecting patency. Thrombus burden describes material deposited on or within the test segment. Bleeding and hemostasis measures show whether inhibition of thrombosis comes with an excessive effect on normal hemostasis. No single endpoint covers all three dimensions of pharmacology.
What the AV shunt model measures
AV shunt thrombosis models place a test device or graft segment in a circuit that exposes it to flowing blood. Contact with a foreign surface, local geometry and the resulting flow field initiate platelet adhesion and activation, coagulation and fibrin formation. The balance among these processes depends on the test article, surface material, device geometry and hemodynamic conditions, as well as the animal's baseline hematologic state.
Historical work established the feasibility of measuring thrombus formation in rhesus monkeys using test devices exposed to blood in arterioarterial or arteriovenous shunts. In one study, thrombus deposited during a defined exposure was quantified gravimetrically; device geometry and material influenced the amount recovered, and the investigators reported more predictable flow in the AV than in the arterioarterial configuration.[1] A related study used the model to compare antithrombotic agents and combined gravimetric thrombus measurement with ultrastructural assessment.[2] These studies provide proof of principle, not a ready-made protocol or universal reference range for a contemporary NHP study.
Treat the model as a controlled, acute thrombogenic challenge built around blood flow and a defined blood-contacting surface. It does not reproduce every feature of spontaneous arterial or venous thrombosis in patients. A device-associated clot endpoint should not be read as a direct proxy for plaque rupture, venous thromboembolism, stroke, myocardial infarction or clinical graft failure.
Three endpoint domains, and why they should stay distinct
1. Flow and shunt patency
Flow is a functional readout of the evolving interaction between thrombus and the test circuit. Depending on the study design, useful summaries include the baseline and serial flow profile, the change from baseline, flow over time, or time to a prespecified flow reduction or interruption. If flow is continuously monitored, the data show when a change begins and whether it is sustained, information that a terminal measurement alone cannot provide.
Flow must be interpreted in relation to the measurement system. Probe placement, circuit resistance, device dimensions, pressure conditions, temperature and technical interruptions affect the recorded signal. A protocol should define the acceptable baseline range, how signal artifacts are handled and what constitutes an interpretable reduction in flow before treatment groups are compared. "Time to occlusion" is meaningful only when the threshold and observation window are prespecified and the model reliably reaches that event under the selected conditions.
A lower thrombus mass does not prove that flow was preserved, and preserved flow does not prove that little thrombus formed. A non-occlusive deposit builds up without an abrupt flow change; a strategically located deposit impairs flow despite a modest total mass. Flow and thrombus burden are complementary endpoints, and neither substitutes for the other.
2. Thrombus burden and composition
Terminal thrombus assessment quantifies the material formed on or within the test segment. Gravimetric measurement is a direct, established approach in the historical primate literature, but the result depends on standardized recovery, handling and weighing. The protocol should specify whether the measured endpoint is wet or processed mass, how residual blood is addressed, and whether results are normalized to a relevant device characteristic such as exposed area or exposure duration. These choices must be consistent across groups and reported in the same way.
Mass alone does not reveal what the thrombus contains or how it formed. Where scientifically justified, imaging or microscopy and targeted assessment of platelet and fibrin components add mechanistic context. Such analyses help distinguish a change in overall deposition from a change in the relative contribution of platelet-rich and fibrin-rich material. Define the method, sampling location and scoring approach in advance, and treat an exploratory morphology readout as exploratory until it has been qualified.
A useful endpoint set links the in-circuit observation to the recovered specimen: flow trajectory during exposure, thrombus amount at the endpoint and, where included, characterization of thrombus composition. That sequence gives a fuller reading than a single terminal weight.
3. Bleeding and hemostasis
Antithrombotic activity and its effects on hemostasis belong in the same analysis. A study typically pairs clinical observations of bleeding, assessment of access or procedural sites, a protocol-defined bleeding measure, and laboratory or pharmacodynamic data relevant to the drug's mechanism. The precise panel follows the test article and the study question rather than a template copied from another species or pharmacology class.
Bleeding time, when used, is a context-dependent measure affected by the method and execution; it is not a direct estimate of clinical bleeding risk. Likewise, routine coagulation tests, platelet counts or platelet-function assays each capture only part of hemostasis and are not equally sensitive to every mechanism. Interpretation is stronger when observations are paired with exposure and pharmacodynamic data, such as the relevant coagulation or platelet pathway assay, rather than treating any single laboratory result as a complete safety conclusion.
Bleeding endpoints also require a standardized approach and an animal welfare review in NHPs. The protocol should state how events are observed and graded, the relevant observation period, criteria for follow-up or intervention, and how procedural bleeding is distinguished from a drug-related signal. A finding of reduced thrombosis with no detected bleeding in a small study does not support a claim of "no bleeding risk."
Build an endpoint package around the pharmacology question
Endpoint selection follows the intended decision. An antiplatelet program usually needs platelet deposition or function, flow, and a hemostasis assessment. An anticoagulant program turns on coagulation-pathway pharmacodynamics and exposure, which are what make a change in thrombus formation interpretable. A surface-modifying device puts the emphasis on material-specific deposition and surface characterization. In every case the endpoints trace back to the mechanism of action and to the limits of the model.
A coherent AV shunt study typically aligns four information streams:
- Functional response: flow behavior and any prespecified patency or occlusion measure.
- Thrombus response: amount deposited, with composition or morphology assessed when justified.
- Pharmacology: test-article exposure and mechanism-relevant pharmacodynamic measurements at interpretable time points.
- Hemostatic context: clinical bleeding observations and selected laboratory or functional measures appropriate to the candidate.
The analysis plan should identify the primary endpoint and distinguish it from supportive or exploratory measures. Randomization, blinded assessment where feasible, control conditions, sample-size justification and predefined handling of technical failures reduce interpretive uncertainty. Where repeated exposure or within-animal comparisons are considered, the design must account for carryover, washout and the biological variability of the selected endpoints.
Sources of variability that change the result
An AV shunt thrombosis result is specific to the experimental configuration. Device material and geometry, exposed surface area, circuit dimensions, flow conditions, exposure duration and peri-procedural management all influence thrombus formation. Biological factors, including platelet count and function, hematocrit, coagulation status, species, individual variability and concomitant medications, also affect the response.
These factors are not merely technical details. If they differ across groups or study periods, they change the apparent treatment effect. Baseline hematology, consistent device preparation and handling, calibrated flow measurement, procedural standardization and complete documentation are part of endpoint validity. A protocol should also define how unexpected interruption of flow, device failure or inadequate thrombus formation will be recorded and treated in the analysis.
The NHP context adds value when the study requires an integrated in vivo assessment of exposure, pharmacology and hemostatic response in a primate. It does not eliminate the need for model qualification. Historical rhesus studies show that AV shunts support studies of thrombus generation and antithrombotic effects, but they do not establish that results transfer unchanged to another species, device configuration, compound class or contemporary study setting.[1,2]
Translational interpretation: a benefit-risk profile
A conclusion that a candidate "reduced thrombosis" says little on its own. The questions are whether the reduction showed up in flow, in thrombus burden, or in both; whether it tracked drug exposure and the expected pharmacodynamic mechanism; and what the hemostasis data show over the same exposure range. Discordance among endpoints is informative, and usually reflects differences in sensitivity, timing, clot location or mechanism rather than a failed experiment.
An NHP AV shunt thrombosis model contributes to preclinical assessment of antithrombotic pharmacology when its configuration and endpoints are matched to the decision at hand. Its strength lies in bringing flow, device-associated thrombus and systemic pharmacology into one experimental context. Its limits matter just as much: it is an acute, configuration-dependent model and cannot alone establish clinical efficacy, clinical bleeding risk or therapeutic benefit in a specific patient population.
For translational teams, the practical priority is to define the question first, then select a flow endpoint, a thrombus endpoint and a hemostasis package designed to be read together. That alignment, not the number of measurements, determines whether the study produces decision-relevant evidence.
Frequently asked questions
Q: What does an NHP AV shunt thrombosis model measure?
A: It evaluates thrombus formation on or within a defined blood-contacting test segment under flow in an NHP. Depending on the protocol, endpoints include flow behavior, recovered thrombus burden, thrombus composition, drug exposure, pharmacodynamic activity and hemostatic effects.
Q: Why measure flow as well as thrombus mass?
A: Flow describes functional patency during the experiment, while thrombus mass quantifies deposited material at the collection endpoint. The measures are related but not interchangeable: thrombus sometimes accumulates without occlusion, and a small strategically located clot affects flow disproportionately.
Q: Does lower thrombus burden mean lower bleeding risk?
A: No. Antithrombotic efficacy and bleeding are related pharmacology domains that must be measured separately. A study should include a mechanism-appropriate hemostasis assessment, and a reduction in thrombus burden alone does not establish an absence of clinical bleeding risk.
Q: Can AV shunt results predict clinical thrombosis outcomes?
A: They provide preclinical evidence about antithrombotic activity in a defined device-and-flow configuration. They do not, by themselves, predict clinical outcomes or reproduce all mechanisms of spontaneous arterial or venous thrombosis.
References
- Mason RG, Wolf RH, Zucker WH, Shinoda BA. A nonhuman primate vascular shunt model for thrombus generation. Laboratory Investigation. 1976;34(6):558-564. PubMed PMID: 819713.
- Mason RG, Wolf RH, Zucker WH, Shinoda BA, Mohammad SF. Effects of antithrombotic agents evaluated in a nonhuman primate vascular shunt model. American Journal of Pathology. 1976. PubMed PMID: 820202; full text.
