Ischemic complications of dermal fillers

Vision loss

As cited in the introduction and the Table, there are at least 211 cases of blindness reported in the English literature as of January 2022, and the actual number is likely many folds higher, since most cases are not published in literature. The presentation of a patient with vascular occlusion of the ophthalmic artery involves blindness and periocular symptoms, usually very soon after filler injection, and can present with concurrent ptosis and ophthalmoplegia^[29,34]. In a recent review paper, 35 out of 39 cases had immediate vision loss symptoms and 2 cases developed symptoms within 10 min^[29]. However, in several cases, the symptoms developed a day after injection^[29]. The etiology of this vision loss is usually due to retrograde arterial embolism into branches of the ophthalmic artery, including the retinal branches^[35-37]. The glabella, temple, and nasolabial folds have vasculature that commonly anastomoses to the ophthalmic artery^[38]. Thus, the most common locations that cause vision changes are the nasal region, followed by the glabella, forehead, and nasolabial folds^[5,39]. The most common occluded vessels are the ophthalmic artery, central retinal artery, branch retinal artery, and naso ciliary artery^[29,30]. Occlusion of the ophthalmic artery is usually secondary to injections in the nose, while occlusion of the retinal artery is secondary to glabellar injections^[30]. Given the various anastomoses in the different facial arteries, there are several injection locations that can cause ocular complications.

A recent study found that filler particles disintegrate into smaller particles immediately after injection, supporting the hypothesis that emboli, rather than a column of filler, cause an obstruction^[40]. Another study further supported this idea by showing that the force required to push a column of filler retrograde was higher than the normal injection force^[41]. However, one study also found that only 0.085 mL was required to fill the supratrochlear artery from the glabella to the ophthalmic and central retinal artery bifurcation, highlighting that even a small volume of filler could cause this complication^[42]. Per the 48 published case reports by Beleznay et al. of filler-induced vision changes, 81% of cases were treated with hyaluronic acid filler followed by calcium hydroxylapatite (10.4%), and one case each was from autologous fat and poly-L-lactic acid (PLLA)^[5]. This is likely secondary to the fact that hyaluronic acid is the most common type of filler injected, followed by autologous fat and calcium hydroxylapatite^[39]. However, visual loss has even been noted to occur secondary to platelet-rich plasma (PRP) injections and PLLA injections^[43,44]. In the 4 reported cases, one woman developed painful vision loss with no light perception in the left eye after PRP injection into the left glabellar region^[43]. Her fundus exam was suggestive of an embolus within the central retinal artery^[43]. At her 1-month visit, she was noted to have a pale optic disc with pigment in the superior temporal region and developed a retinal detachment 8 months after her initial injection^[43]. In one reported case of blindness secondary to PLLA, a patient received an injection in the left periorbital region and reported immediate pain in the left eye as well as blurring of vision^[44]. On exam, the patient was found to have a decrease in vision to light perception with projection, a fixed and dilated left pupil, as well as ophthalmoplegia and ptosis^[44]. While the ophthalmoplegia and ptosis improved over time, the patient’s vision still declined^[44].

Although irreversible damage to the retina has been previously studied to occur within 90 min, a recent study found that retinal infarction can happen as early as 12-15 min after complete occlusion^[45,46]. Given the low incidence of cases, no high-level studies currently exist that allow for a clear management recommendation for this devastating complication^[47]. Ocular physical maneuvers, including compression and paracentesis, are a rapid intervention^[47,48]. An ocular massage can be performed with manual firm compression to the globe in 10-20 second intervals followed by a sudden release. This compression can be performed with a Goldman lens or trans palpebral with 2 fingers^[47]. The goal of these maneuvers is to lower the intraocular pressure, dilate the occluded artery and allow for the migration of the emboli to a peripheral vessel, preserving central vision^[47]. Rebreathing in a brown paper bag for 10 min every 30 min can also increase CO₂ and cause vasodilatation^[49]. Methylprednisolone and other intravenous steroids can be used to decrease the retinal edema caused by damage to the cells^[11,47]. Intraocular pressure reduction can be achieved with timolol, mannitol, and acetazolamide to restore retinal vascular flow and avoid visual loss^[47]. Lastly, hyperbaric oxygen (HBOT) can be used when available, as it provides a subjective relief and can also work to improve plasma oxygen concentration and dilate the retinal arteries^[47,50]. While nitroglycerin paste has been considered, questions remain on whether it is able to penetrate the deep orbital vessels^[47]. One study with a rabbit eye model found that there was no improvement in perfusion with nitroglycerin paste, and the veins were also found to have a more congested appearance^[51]. Additionally, dilation of the arterioles could push the product further into the smaller arterioles and capillaries^[51]. The use of hyaluronidase can also be considered, especially when the filler is HA-based, within 60-90 min after visual loss^[47]. While hyaluronidase has been noted to also be administered outside the 90-min window, the newest cases have not been able to show significant improvement in visual acuity.

There has been some interest in the use of HBOT in the treatment of ophthalmic artery occlusion. HBOT is defined as an intervention in which an individual “breathes near 100% oxygen intermittently while inside a treatment chamber at a pressure higher than at sea level pressure (> 1 atm)^[52].” The main goal is to increase the amount of oxygen dissolved in the plasma. Increased dissolved oxygen works to improve the diffusion distance, supporting oxygen-dependent processes that do not get enough arterial blood supply^[53]. One report discussed the case of a 31-year-old woman who received HA filler at the nasal dorsum and developed immediate vision loss in the right eye^[54]. After being diagnosed with central retinal artery occlusion, she received a multitude of treatments including high-flow O₂, ocular massage, steroids, and blood thinners^[54]. She also received daily 90-min therapy of HBOT at 2.5 atmospheres absolute (ATA) or 253 kilopascals (kPa) for 3 weeks. However, the patient had no improvement in the vision in the right eye, with her visual acuity remaining at no light perception^[54]. In another case, a 49-year-old woman was also found to develop a central visual defect in her right eye after the injection of poly-L-lactic acid (PLLA)^[55]. Her visual acuity was found to be 20/200 in the right eye. She received ocular massage, brimonidine, and HBOT (unknown pressure) twice daily for 5 days, but did not have any improvement in her vision^[55]. A 41-year-old woman who had vision loss in her right eye after a forehead filler injection received retrobulbar hyaluronidase and 2 h of daily hyperbaric oxygen therapy, among other treatments^[56]. This patient had recovery of vision from no light perception to hand motion^[56]. It appears that an optimal HBOT protocol is daily use for several days or weeks after the onset of blindness. The typical HBOT protocol for compromised grafts and flaps is 2.0 to 2.5 ATA for 90 to 120 min twice a day, and it appears that preliminary case reports are following that protocol^[52]. While the use of HBOT has been shown to have some effect on skin vascular occlusion, discussed later in this review, there is still the question of whether HBOT has the same effect on visual loss from occlusion. It is worth noting that data from studies solely looking at artery occlusions outside of emboli from facial filler have documented some improvement with HBOT^[57-60]. The disease processes in these studies looked at idiopathic CRAO and Factor V Leiden mutations^[57-60]. However, at this stage, the data for HBOT for ophthalmic artery occlusions and central retinal artery occlusions secondary to facial filler emboli remain preliminary and inconclusive.

The use of hyaluronidase through a retrobulbar injection is currently being debated and further studied. A 2018 case report discussed a retrobulbar injection of 450 IU of hyaluronidase, which completely restored a patient’s vision after receiving HA filler injections near the infraorbital neurovascular bundle^[61]. However, this report did not document an objective visual acuity before and after hyaluronidase treatment. Another case study also showed improvement from light perception to hand motion and eventually full recovery after 2 injections of 900 IU retrobulbar hyaluronidase, 4 and 5 days after filler injection^[62]. Given that these cases are anecdotal and not controlled studies, it is unclear if an improvement in vision would have occurred without retrobulbar hyaluronidase. Additionally, other reports have not shown improvement or been as definitive as to whether this treatment is a viable solution^[63,64]. An animal model study showed that extravascular injection of hyaluronidase was not able to penetrate the vascular lumen or re-perfuse occluded auricular arteries^[65]. This is supported by Hwang et al., in which 1000 IU of retrobulbar hyaluronidase within the ophthalmic artery 30 min after occlusion in rabbits was unable to relieve the obstruction^[66]. However, in contrast, Lee et al. injected hyaluronic acid into the ophthalmic artery, confirmed ischemia with fundus photography, and then injected retrobulbar hyaluronidase^[67]. While initial experiments with 1500 IU did not show any improvement, 3000 IU of hyaluronidase 5-10 min after occlusion showed an improvement in perfusion, possibly secondary to higher dosing and faster treatment^[67]. Questions were raised about the methodology in the Lee et al. study regarding the lack of documented electroretinogram testing prior to treatment with hyaluronidase to detect a complete occlusion^[67,68]. Hwang et al. found that the fundus photographs seemed to support a branch retinal artery occlusion rather than a complete retinal artery occlusion, possibly indicating only partial vision loss prior to treatment^[68]. Additionally, the authors discuss that a dose of 3000 IU could lead to a compartment syndrome with readily available preparations (since the powdered form is not widely used in the US) and injecting within 5-10 min of occlusion would prove to be very difficult in the clinical setting^[68]. Another in-vitro study showed that hyaluronidase could not cross the dural sheath of the optic nerve, which can prevent access to the central retinal artery^[69]. Ugradar et al. also found that hyaluronidase was not able to reduce the particle size in a substantial manner, creating particles that were still greater than the size of the vessel^[70]. These in-vitro studies highlight that physiologically, it is difficult to replicate how retrobulbar hyaluronidase can effectively relieve filler obstructions. This translates to what has been seen clinically as well, in that some patients have a full recovery of vision with retrobulbar hyaluronidase while other cases do not^[29]. An additional literature review concluded that the efficacy of retrobulbar hyaluronidase was not clear and that there was not enough evidence to support retrobulbar hyaluronidase given the inherent risks^[48]. This treatment has a level of evidence V, which confers a grade D recommendation from the American Society of Plastic Surgeons^[48,71].

A more novel way to treat blindness from filler complications may be to consider it a stroke, because the optic nerve is part of the central nervous system. Activating a stroke protocol may be prudent. Baley-Spindel et al. noted the occlusion is likely composed of both HA gel and red thrombi; therefore, a combination of hyaluronidase and alteplase yielded the best results in clearing HA gel thrombi in their rat model^[72].

Ophthalmoplegia/Double vision

While vision-threatening complications are better known, facial fillers have also been seen to cause ophthalmoplegia, as noted in Table 1. A recent study showed that 50% of patients with occlusion of the ophthalmic arfpresentation is not always as cleartery also presented with ophthalmoplegia^[73]. Another study found that 15 (71%) of 21 patients over a 9-year period with artery occlusion also developed ophthalmoplegia at initial presentation, with an average of 2.8 rectus muscles involved^[74]. The mechanism behind this is theorized to be from ischemia to the cranial nerves or extraocular muscles^[74,75]. However, a recent report of 2 isolated ophthalmoplegia cases illustrated that this side effect could also be secondary to an inflammatory response^[76].

Management of ophthalmoplegia, if a hyaluronic acid (HA) filler is used, can be performed with 1500 IU of hyaluronidase subcutaneously around the site of injection, and in the case reported by Bae et al., hospitalization was required^[77]. Even though the ophthalmoplegia is resolved in some cases, sometimes the ocular misalignment persists, requiring strabismus surgery^[76]. In the 2 case reports where an ischemic etiology was not suspected, one patient improved after Medrol Dosepak and aspirin 81 mg, and the other did not improve as she presented 3-4 months after the complication^[76].

Soft tissue necrosis

The presentation of a patient with vascular occlusion is often characterized by pain disproportionate to the injection, along with blanching^[78]. This is followed by livedo reticularis secondary to venule swelling^[79,80]. However, in a clinical setting, the presentation is not always as clear, with pain often not accompanying these events^[81]. For example, a net-like reddish/blue appearance may be hard to distinguish from bruising or simple erythema, but a delayed capillary refill would support ischemia. A recent study also showed that in a survey of 52 injectors, 62% had reported one or more intravascular injections, highlighting the frequency of these events^[81].

The etiology of vascular complications involves arterial compromise causing tissue anoxia and progression to necrosis^[15]. As the filler is being injected, it may flow in either direction in the vessel, and lead to an obstruction of the blood supply^[15]. Given that fillers have inherently different properties in viscoelasticity and cohesivity, the outcomes have different severity^[3,15]. Case reports with polymethylmethacrylate (PMMA) have had complications showing more dramatic skin necrosis compared to other fillers^[82-84].

The glabella, nose, and nasolabial folds are at higher risk because they depend on a single arterial branch^[13]. Out of all the cases of vascular necrosis, nasolabial fold injection was associated with the highest number of cases, followed by injections into the nose^[85]. Nasal necrosis is likely secondary to the dorsal nasal artery having variable anatomy, occurring only 34% of the time as a pair of arteries, with other variations including a single large dorsal nasal artery or in random distribution^[86]. Given the potential for embolization of facial fillers, vascular complications also have been seen to occur far from the injection area in the mid and lower face. The vascular supply of the internal nose arises from branches of the superior labial artery, the sphenopalatine artery, the posterior ethmoidal artery and the anterior ethmoidal artery^[87]. After an injection of HA above the anterior nasal spine and nasal bones, a 42-year-old female developed gingival necrosis of the right upper incisor, partial lip mucosa necrosis, and an exophytic palatal lesion^[87]. The gingival necrosis was likely related to embolization of the septal branches of the superior labial artery and compression of the distal arteries from the septal branch of the posterior ethmoidal artery^[87]. In one author’s personal experience, she received an injection of hydroxylapatite in her left cheek, in which she noted that the needle was directed somewhat tangentially. After injection, she immediately developed blanching and mottling, which later took on a reticulated appearance due to the involvement of the infraorbital artery. She did not receive hyaluronidase at that time. The following morning, the mottling of her skin extended towards her lower eyelids and the left nasal bridge [Figure 2]. On day three, she developed pustules, necrosis of the skin, and eventually left permanent scarring [Figure 3]. Her erythema of the gums was suggestive of infraorbital artery involvement [Figure 4]. Injections into the infraorbital artery may be performed from the cheek augmentation, with the subsequent embolization extending to the facial artery below.

Figure 2. Mottling of the skin the morning after cheek injection, extending from the lower eyelid to the upper lip and part of the nose due to involvement of the infraorbital artery.

Figure 3. Formation of pustules on the nose and a full thickness defect near the nasolabial fold 3 days after injection due to involvement of the facial artery.

Figure 4. Erythema of the gums indicating involvement of the infraorbital artery.

Skin necrosis has also been seen to occur from filler injected into the lower face. One patient, after receiving dermal fillers in the lip to correct an atrophic scar, presented with pain and blanching of the upper lip with a blue tinge 2 hours after her initial injection^[88]. The authors did note that the necrosis may have been secondary to an alteration of the blood vessel course from scarring already present in the area^[88]. Injections in the lower lip may need to be cautious of the inferior labial artery, which has been studied to have variation in its dominant arterial sources [Figure 5]^[89].

Figure 5. Representative figure of the main arteries of the lower face and chin, with depiction of the variation of arterial sources of the inferior labial arteries. Figure reproduced from Tansatit et al.^[89].

Skin necrosis with chin injections has also been reported^[90,91]. In one case, the patient developed numbness on the right side of her tongue during an injection of her chin and was found to have an obstruction of her deep lingual artery^[90]. In another case, after chin augmentation injection, a patient developed pain with swallowing as well as livedo reticularis and mottling from the mental crease to the upper cervical area 10 min after injection^[91]. Pain with swallowing can be explained by ischemia in the submental branches supplying the digastric, mylohyoid, and platysma muscles^[91]. It is worth noting that the main blood supply of the chin comes from the ascending mental artery, branching off the submental artery^[33]. If that ascending mental artery is cannulated, the filler can travel retrograde to the submental artery and facial artery, causing infarction to the lips, nasolabial fold, nose, and paranasal skin^[33]. Additionally, if filler travels through the submental artery and crosses anastomosis to the sublingual artery, it can also cause necrosis of the root of the tongue and floor of the mouth through the involvement of the dorsal lingual and sublingual arteries^[33].

If vascular occlusion is suspected with pallor and blanching, the injection must be stopped immediately^[13,78,92]. Aspiration of the product should be attempted before withdrawing the needle^[13,78]. With the goal of improving blood flow, warm compresses and a massage of the area should be done^[13,76,90]. Regardless of filler type, hyaluronidase should be injected as it has been shown to reduce edema^[93,94]. Aspirin should also be started to limit clot propagation and platelet activation and be given with an antacid^[13,92]. In one case report, low level light therapy (LLLT) was used with the intention of reducing pain and inflammation, and improving tissue repair and regeneration^[77]. The goal of LLLT is to use photons at a non-thermal value to change biological activity^[95]. Through the enhancement of specific enzyme activity, LLLT has been shown to activate intracellular signaling pathways and transcription factors involved with cell proliferation, survival, and tissue repair^[95]. Nishioka et al. have also shown that LLLT therapy increases skin flap viability in rats, with the percentage of necrosis area of the flap decreasing in the LLLT group^[96]. Sildenafil, tadalafil, and vardenafil can also be used to relax smooth muscles, dilate blood vessels, and increase blood flow^[97]. The use of nitroglycerin has not been fully defined. Although nitroglycerin paste has innately vasodilatory properties, an animal study showed no improvement in perfusion and raised the question of whether arterioles dilation could cause further propagation of filler and worsen the ischemia^[51]. Given its inherent risks of headaches and hypotension, van Loghem et al. did not make a clear recommendation on whether to incorporate nitroglycerin paste^[13].

Hyperbaric oxygen has also been reported to be an effective adjunct treatment for vascular occlusion. A 32-year-old female who developed vascular occlusion after HA filler on her nose received 1 month of biweekly 90 min HBOT sessions at 2.4 ATA, after which her nose showed improvement in vascularization^[98]. Another report also remarked on a 37-year-old female who developed ischemic changes to her face after an injection on her proximal temple^[99]. After receiving 6 treatments twice daily (2 treatments at 3.0 ATA followed by 4 treatments at 2.4 ATA), the patient showed improvement in ischemic discoloration^[99]. Two cases, one involving a 46-year-old man who received poly-methylmethacrylate and calcium hydroxylapatite and a 40-year-old white woman who was treated with hyaluronic acid, showed improvement in their skin necrosis after HBOT^[94]. One treatment lasted 14 days, while the other lasted 2 days^[94]. Thus, while HBOT remains an inconclusive treatment for vascular occlusion of the ophthalmic and retinal artery, it has been shown to potentially have more benefit in skin necrosis treatment.

Recently, a new protocol has been released in which high dose pulsed hyaluronidase is used for vascular adverse events. For low volume events (0.1 mL or less of HA filler), 450 IU of hyaluronidase is used in a single area, for an area half of an upper lip^[100]. If the nose has involvement, 900 IU of hyaluronidase should be used^[100]. The dosing should also be hourly rather than the traditional daily dosing to maintain high concentrations in the ischemic zone^[100]. The injections should be given at injection sites, and if there are distal sites that appear ischemic, one should consider injecting distal sites of ischemia^[15]. Delorenzi et al. have remarked that hyaluronidase is able to diffuse through an arterial wall, so the most important area of injection would be areas that show ischemia^[15]. However, the article did comment on a case in which the patient did not initially improve with hyaluronidase injection into the ischemic tissue, but showed improvement after hyaluronidase was injected into the affected artery^[15]. Thus, further studies may be needed to elucidate the most effective area for hyaluronidase injection immediately after vascular occlusion.

Cerebral infarction

Filler injections can lead to very severe complications. Cerebral infarction secondary to vascular occlusion has been noted as a complication of filler injections. A review article of 44 cases showed that 8 patients (18.2%) had CNS involvement, including upper limb weakness, acute infarction, or hemorrhage^[29]. A 20-year-old female who presented with non-improving vision loss in her right eye was found to have multifocal infarcts in her parietal lobes^[101]. This patient received injections in the glabella, an area known to cause combined ophthalmic and cerebral complications^[101]. Anatomically, the supratrochlear and supraorbital arteries were injected with enough force that the filler traveled retrogradely to enter the cerebral circulation via the Circle of Willis^[101]. Another study involved a 39-year-old female who presented with vision loss in the left eye after filler injection into the glabella^[102]. This vision loss was concurrent with ptosis and total ophthalmoplegia^[102]. The next day, an MRI found several cerebral infarctions that were embolic in nature^[102]. After one week, the infarction transformed into a parenchymal hematoma, after which the patient received methylprednisolone^[102]. While her limb weakness improved, she continued to have a right arm monoparesis^[102]. Another case involved a 40-year-old female who developed a cerebral infarction after nasal augmentation^[103]. With a GCS of 4 and on mechanical ventilation, the patient developed gastric ulceration, pulmonary infection, respiratory failure, and cerebral herniation, dying 6 days after the filler injection^[103]. For this patient, the theory was that there were several small HA particles in the capillaries from this filler injection^[103]. The authors recommended immediate thrombolysis within the 12-h window of functional impairment (90 min for concomitant ophthalmic arterial occlusion) and consideration of decompressive craniectomy^[103].

Injection techniques

There are several injection techniques that have been studied to help prevent vascular occlusions. Aspiration before injection (with an unprimed needle, especially if the filler is thick or sticky) and using lower volumes of the product (0.1 mL) can help indicate that the needle has not entered a blood vessel and reduce the severity of complications if a blood vessel is entered^[7,17]. Sometimes aspiration can give you a false sense of security, as the filler may be too thick to detect reflux of blood when in a vessel. If injecting in the areas of the supratrochlear artery, supraorbital artery, and dorsal nasal artery, compression of the vessel pathway can help prevent retrograde flow^[104]. The use of a reversible HA filler will allow for treatment with hyaluronidase if a reversal of vascular occlusion is needed^[7]. Given that the pressure of the filler injection can cause retrograde flow, injecting at a slow pace can help in preventing complications^[7,105].

Needle size has also been seen to have an impact on vascular obstruction. While a smaller needle may improve the precision of the injection, it may increase the likelihood of penetrating the vessel wall rather than a larger bore needle which would roll on the side of the artery^[15,104]. Additionally, more pressure is required to inject through a smaller bore needle which could lead to more pressure into a vessel, should vasculature be entered. A recent study also found that the use of microcannulas had 77% lower odds of occlusion compared to needle injections, due to the fact that blunt tips could avoid piercing vessel walls^[106]. However, small bore microcannulas still have the potential to penetrate the facial artery with a small amount of force (0.23 kg for a 27G cannula)^[107]. Epinephrine may also help with vasoconstriction, but may make it more difficult to distinguish an early manifestation of necrosis^[104]. Thus, administering local anesthesia without epinephrine may be considered^[104,108].

Imaging

Mapping vasculature to prevent intravascular filler injection

Various studies have now started incorporating ultrasound as a method to map vasculature and reduce the incidence of vascular complications seen with filler injections. Doppler ultrasound has been seen to detect anatomy, such as the facial artery lateral to the nasolabial fold, along with its different anatomical variations^[109]. One other study commented that although ultrasound may be difficult to manage in conjunction with injection, a new technique combining 3D time of flight magnetic resonance angiography (3D-TOF MOTSA MRA) and infrared (IR) facial heating could visualize the following arteries: facial, angular, superior labial, inferior labial, lateral nasal, dorsal nasal, supratrochlear, supraorbital, and superficial temporal^[110]. These images could be acquired on a 1.5 or 3 Tesla (T) system with a head coil, with an additional surface coil on top, to improve signal reception^[110]. The authors showed that these MRA images could be projected on the patient’s face before injection, albeit without the same 3D depth aspect^[110]. The main benefit of this method was visualization of the facial arteries in a non-invasive and contrast-free manner^[110]. However, the cost of MRI in these individual countries may need to be considered before the widespread adoption of this technology. Our team has also looked at the utilization of 7T MRI and shown its ability to depict small orbital and eyelid structures, as well as the orbital branch of the infraorbital artery (article in press).

The use of ultrasound has been adopted at various points in the filler injection procedure. Firstly, ultrasound has successfully determined that the needle or cannula is positioned in the correct plane and not within or near a vessel^[111]. Schelke et al. discussed how the use of doppler ultrasound can help distinguish vessels as well as the direction of blood flow [Figure 6]^[112]. Rocha et al. recommended that once the needle or cannula is determined to be in the correct location, the filler can be injected without aspiration^[111]. After the injection of filler, doppler ultrasound can again be used to confirm the vascularization of the area^[111]. As seen in Figure 7, the placement of filler deposits, such as hyaluronic acid, can be distinctly visualized on ultrasound to confirm correct depth and placement.

Figure 6. Localization of the artery with ultrasound. Figure reproduced with permission from Schelke et al.^[112].

Figure 7. Multiple deposits of hyaluronic acid filler, as represented by the two anechoic deposits and one hypoechoic deposit. Figure reproduced with permission from Schelke et al.^[112].

Diagnosing areas of ischemia from filler complications

Imaging can also be used in the early diagnosis and treatment of filler complications. Another study evaluating laser doppler imaging (LDI) found that LDI was able to accurately delineate a hypo-perfused area to help target hyaluronidase treatment^[113]. Color Doppler flow imaging (CDFI) was able to successfully detect retinal artery occlusions and ophthalmic artery caused by filler injections by showing decreased retrobulbar blood flow^[114]. Doppler ultrasound may be able to be used to detect the lack of perfusion, and in rare cases, an ischemic vessel may be able to be injected with hyaluronidase under ultrasound guidance. Schelke et al. reported on a case of how, upon crusting under the lip after upper lip augmentation, 150 U of hyaluronidase was injected with ultrasound guidance^[112]. A case in which ultrasound was used to inject hyaluronidase into a visualized filler deposit is seen in Figure 8. In another case, after a patient presented with mottling of her chin following filler injection, ultrasound was used as guidance for hyaluronidase injection [Figure 9]. Ultrasound imaging appears to be valuable in helping the injector decide where to place the hyaluronidase. However, this may be very difficult to do, and it is unclear how much an ultrasound may help or distract from the process of trying to cannulate a vessel.

Figure 8. Needle inserted into filler deposit under ultrasound guidance. Figure reproduced with permission from Schelke et al.^[112].

Figure 9. Using ultrasound (Clarius Portable Ultrasound L20) in a clinical setting to localize filler placement causing skin ischemia and injecting hyaluronidase (see Supplementary Video 1 and Supplementary Video 2)

Magnetic resonance imaging (MRI) has also been studied as the primary mechanism to evaluate for infarctions after ischemic complications. Many of the case reports that were reviewed showed that an MRI or magnetic resonance angiography (MRA) was used to evaluate intracranial infarctions and embolisms^{[101,102,115]}. Additionally, the use of MRI is finding an increasing role in the detection of other non-ischemic complications associated with fillers such as inflammation, foreign body granulomas, and filler migrations, which we will expand upon in another review paper^[116]. While the use of imaging in this field is still relatively recent, it represents an area for further development and utilization.