Neurological complications of foam sclerotherapy: fears and reality

The issues can be summarized into only one sentence: what are the neurological risks of foam sclerotherapy (FS)?

We know that acoustic contact emboli were detected in the right side of the heart in patients after FS of the great saphenous vein. ^1,2 These echoes certainly represent bubbles but probably other particles as well.

In patients with a patent foramen ovale (PFO), signals were detected in the left heart and the middle cerebral artery (MCA) by transcranial Doppler (TCD). ^1–4

The clinical significance of microemboli is not clear – Gibson et al. ⁵ reported a series of 82 patients assessed by TCD. A diffusion-weighted magnetic resonance image (MRI) was performed at one and 28 days after FS. MCA signals were demonstrated in 57 patients but no cerebral lesions were detected.

Morrison ⁶ presented a series of 59 patients treated with FS and monitored by TCD of the MCA. High intensity transient signals were recorded in 37% of symptomatic patients but, interestingly, also in 63% of asymptomatic patients.

Among the neurological disturbances, we must distinguish visual disturbances and cerebro-vascular events.

Visual disturbances

Visual disturbances (ViDs), described after injection of sclerosing agent in the liquid form, appear to be more frequent after FS. ⁷ Their frequency of occurrence is estimated around 1.5%. ^8,9 They are described by patients as reversible symptoms including positive features (e.g. flickering lights, spots, lines or scotoma) or negatives features (i.e. loss of vision). They could involve one or both eyes. In a study carried out with the collaboration of neurologists, ¹⁰ based on a clinical assessment combined with a brain diffusion-weighted MRI, we demonstrated that ViDs correspond to migraine with aura (MA) and are not transient ischaemic cerebro-vascular events.

ViD can be associated with paresthaesia and dysphasic speech disturbance depending on the extension of the cortical spreading depression (CSD) in the cerebral cortex. CSD is the pathophysiological correlate of MA. ¹¹ Two pathophysiological mechanisms of MA occurring after FS are possible and might coexist, which could explain the extent of some reported attacks in patients with a large PFO: a release of endothelin and microembolization with a decrease in cerebral oxygen saturation, both triggering CSD and, thereafter, MA attack.

Endothelin has been demonstrated as being a trigger factor for MA. ¹² Frullini et al. ¹³ have studied systemic endothelin 1 levels after liquid and foam sclerotherapy with lauromacrogol in rats. Whereas endothelin 1 levels did not change significantly in the control and the liquid sclerotherapy groups, a significant increase was identified after foam sclerotherapy at one and five minutes.

Nozari et al. ¹⁴ have demonstrated, in mice, that microemboli of microbubbles of air, polystyrene microspheres or cholesterol crystals into carotid artery could trigger CSD without requisite tissue damage. In this study, air microemboli caused CSD in all mice. CSDs were preceded by global or regional hypoperfusion with a close correspondence between the magnitude and duration of the cerebral blood flow reduction and the appearance of CSD. Despite extensive histopathological evaluation (all animals were examined), no ischaemic infarct was detected in brains after microemboli of air.

Caputi et al. ¹⁵ performed contrast-enhanced transcranial Doppler (ce-TCD) with air-mixed saline in 159 patients with MA. An occurrence of a typical MA attack was overall observed in 12 out of 159 patients, but arose only in PFO positive ones, immediately after ce-TCD (12/79 = 15.2%).

These data reinforce the hypothesis that ViDs occurring after FS correspond to MA and are not transient ischaemic cerebro-vascular events.

Cerebrovascular events

While millions of FS sessions have been performed, only a few cases of stroke related to paradoxical air embolism, ^16–18 all with complete or near complete recovery, and a few cases of transient ischaemic attack (TIA) ^9,17,19–23 have been reported. No death or stroke with significant after-effects related to FS have been reported to date.

One can object that not all cases of FS complications have been reported in the literature. This remark is relevant, but it applies with all methods of treatment, with surgery or endovenous thermal ablation. Furthermore, on the basis of insurance reports, no serious complications related to foam injection and leading to litigation have been reported in France.

We must remember that several cases of stroke related to liquid sclerotherapy ^24–27 and saphenous vein stripping ²⁸ have been reported. One case of stroke following endovenous laser ablation treatment has been recently described. ²⁹

The clinical description of the neurological events related to FS does not always make it possible to assert the diagnosis of stroke. In two cases, ^9,22 the patients experienced a speech disturbance combined with paresthaesia ⁹ or ViD. ²² The diagnosis of MA could be discussed, all the more as the MIRs were normal in both patients. In other reported cases, ^18,20,21 the onset of symptoms was delayed after the injection of foam, making the relationship uncertain.

In some cases of stroke, ^16,17 foam volume or quality or both may be involved. In the first case the patient was treated with 20 mL of polidocanol foam 0.5%. ¹⁶ In the second case, ¹⁷ 10 mL of foam (neither indication of product nor concentration was given) were injected over a 20-minute treatment period to treat reticular and spider telangiectasia.

In one case reported by Ma et al., ¹⁸ the patient had suffered a stroke at the age of 39. Ultrasound-guided FS is a safe and effective treatment for varicose veins. It is used worldwide today. However, as in all treatments, it is necessary to evaluate the ratio benefit/risk before treating a patient with FS.

The type of gas (air or more physiological gas) to prepare foam is a controversial topic.

According to Morrison's studies, the frequency of occurrence of ViD seems to be reduced by substituting CO₂ ³⁰ and CO₂-O₂ ³¹ for air, but the differences between the air-foam group and the other two groups were not statistically significant despite the fact that large volumes of foam were injected (the average volume in each of groups was larger than 22 mL).

CO₂-O₂-based foam could be of interest when large volumes of foam are required (a study is ongoing in Germany). We reported one case of TIA or MA following the injection of foam made with O₂. ⁹

Considering the high prevalence of PFO, which is estimated to be around 30% in the adult population, with a rate up to 70% including all kinds of right-to-left shunts, ³² the risk of stroke following FS appears to be very low. According to the opinion of experts, ³³ screening for right-to-left shunts is not necessary before FS.

Though strokes are exceptional, their prevention must be the physician's main concern.

It is based on the quality of foam and the injected volume. Injection of large volume of foam remains controversial. Most physicians recommend limiting the volume, with a maximum volume per session of 10 mL, ³³ 12 mL (Venous Forum of the Royal Society of Medicine) or 15 mL (Australasian College of Phlebology). They recommend the patient avoid a Valsalva manoeuvre following the injection of foam.

Finally, some measures, such as leg elevation, remaining supine for five minutes or occlusion of the saphenofemoral junction, have been suggested as methods to prevent foam migration but their efficiency is not established. ³⁴