Although the clinical consequences of small air bubbles introduced during endovascular procedures for cerebrovascular lesions are not yet well understood (1-7), the occurrence of any air embolism may be hazardous. Even catheter flushing with saline can result in air embolisms (8). Air embolism during cerebral angiography has also been observed using transcranial Doppler (9). Modification of the flushing technique and the use of heparin and air filters may, however, reduce the occurrence of these embolic phenomena (8, 10). Air-free endovascular procedures, however, require a deliberate effort to minimize the introduction of air bubbles, starting with the use of air-proof devices.

It is surprising that diffusion-weighted images obtained after embolization of unruptured intracranial aneurysms show the same or a more severe degree of microembolic lesions as after carotid stenting procedures, the latter of which are associated with a much higher risk of thromboembolism due to the nature of the procedure (11-14). To explain this, we moved our focus from thromboembolism to air embolism. We therefore assessed the air-proof qualities of the detachable embolization coils we use for the endovascular treatment of cerebral aneurysms. We designed an in vitro system to monitor the emergence of air bubbles during detachable coil delivery.

Air bubbles emerged in 23 of the 30 tested coils (76.7%), most frequently right after the removal of the pusher wire. The total volume of air ranged from 0 to 23.42 mm³ (median: 0.16 mm³) (Table 1). A significantly greater amount of air emerged during the manipulation of Target detachment coils than during that of any other coils (p = 0.0004) (Figs. 2, 3, 4, and supplemented online video clip).

In general, so-called 'cerebral air embolism' is a serious clinical hazard, causing cerebral arterial occlusion and leading to increased intracranial pressure, cerebral infarction, and/or changes in the blood-brain barrier (BBB) (15-18). Several animal studies have shown that 0.5 mL/kg of intracarotid air was sufficient to cause permanent brain damage (19-21). It has been reported that the degree and type of brain injury are largely dependent on the size of these air bubbles (22). A microbubble may cause focal infarction if it is large enough to occlude the arteriole. Although rapid clearance of embolized air bubbles was observed in an animal study (23), smaller bubbles that passed through small arterioles could irritate their endothelium, causing transient BBB breakdown (15, 16). In rabbits, the intracarotid injection of 25 microliters (µL) of air bubbles resulted in marked dilatation of the affected pial arterioles, which persisted for 90 minutes and was directly associated with decreased regional blood flow and depression of neuronal function (15).

Fortunately, the volume of air we could observe with a single detachment coil system in our experiment was much lower than the amount of air found in those serious situations mentioned above. The volume varied according to coil type, with most bubbles having volumes smaller than 25 microliters. However, these bubbles cannot be ignored since, in most patients, multiple coils are implanted, retrieved and reinserted during endovascular treatment of cerebral aneurysms even in a typical procedure.

Every manufacturer recommends profuse back flushing of the coil delivery sheath before inserting the coils into the microcatheter, in order to remove any air bubbles in the sheath lumen. Even more, the Target coil system requires profuse flushing of the dispenser coil tube before the introducer sheath can be removed from the dispenser. As described earlier, we strictly followed the instructions provided by each manufacturer. Then what could be the possible mechanism of air bubble formation during and after delivery of the coil through the microcatheter? What could be the source of these air bubbles?

As Han et al. (24) already observed, the electrolytic mechanism of GDCs could be one of the sources of air bubble formation, despite their small volume. According to our experiment, however, most of the air bubbles were likely due to air introduced by the coil pusher, which was probably why most of the air bubbles were observed immediately after the removal of the coil pusher. It is of note that the volume of air bubbles from the Target coil system was significantly greater than that from any other coil system. We paid particular attention to the recently modified feature of the Target coil system, its pusher wire. The manufacturer modified the prototype GDC coil pusher to provide extra softness and flexibility of the distal segment of the pusher. They replaced approximately 45 cm of the distal segment of the prototype GDC coil pusher with a very flexible wound wire type pusher, which has a serrated irregular surface due to a wire wound around the primary pusher wire. We assume that the change of the surface characteristics could be the source of air bubble introduction. With this we can explain why the total amount of air volume between GDC coils and Target coils is significantly different. The air volume generated during electrolytic detachment was the same for both GDC coils and Target coils, while the volumes observed after removal of the coil pusher differed significantly. Successive tests should be carried out to verify our speculation. Furthermore, modifications by the manufacturer of the design and pusher material are required to resolve this potentially dangerous problem.

The major limitation of our in vitro experiment was the relatively low pressure of our circuit, which was approximately 30-40 mm Hg. This may exaggerate the total air volume due to the inverse relationship between air volume and applied pressure. This test should therefore be repeated under higher physiologic pressure, 80-120 mm Hg, which can be achieved by raising the height of the saline reservoir bottle. However, we monitored air bubbles at room temperature, which is lower than body temperature; this may have compensated for the relatively low pressure we used.