Fluid flow through intravenous cannulae in a clinical model.

Background: Predicting flow through an IV cannula is useful to clinicians if changes in flow are required and to guide selection of cannula. We sought the usefulness of manufacturers' quoted flows in predicting actual flow and to characterize that flow.

Methods: We built a vein model and inserted cannulae between 14 and 20-gauge. In the first experiment, we compared the manufacturer's quoted flows using deionized water, Hartmann's solution and Gelofusine. In the second experiment, we varied the pressure feeding the cannula and measured the resulting flow.

Results: Flow through a cannula is not a simple ratio of the manufacturers' quoted flow rate, even controlling for fluid type and feeding pressure. Flow is neither fully laminar, nor fully turbulent in the range of rates we have measured and in the International Organization for Standardization test. The Reynolds number is often below 2000.

Conclusions: Flow through cannulae is not laminar at the upper range of clinically used flows, therefore Poiseuille's law is not useful in predicting flow and the effect of changing radius is less than commonly believed. The quoted maximum flows are also not useful. There are many conditions for laminar flow apart from Reynolds number. Further work would determine useful predictors of flow.