Commentary on: Efficacy of Retrobulbar Hyaluronidase Injection for Vision Loss Resulting from Hyaluronic Acid Filler Embolization

The problem of sudden loss of vision associated with cosmetic facial filler treatments is a rare event that strikes fear in most of us. As of this writing, there are no known effective treatments for patients with blindness, although several authors have promoted retrobulbar injections of hyaluronidase.1,2 As far as I am aware, patients with total blindness − no light perception (NLP) − at presentation have not responded well to any treatment protocol. In contrast, partial success is not uncommon in patients with only partial visual field obstruction at presentation (in short, NLP seems to be a very bad prognostic indicator). It is not clear whether the treatment was responsible for the observed improvement, since there are no control groups in these scattered case reports, and some patients with partial defects improved spontaneously with only supportive care. This article consists of a report of four cases of filler associated vision loss treated (unsuccessfully) with retrobulbar hyaluronidase (HYAL) and a thoughtful review of the literature.3 The authors argue that retrobulbar HYAL is ineffective. All facial filler treatments carry some risk of vascular adverse events (AEs). We also know that there are several areas of higher risk, namely the glabella, nose, and nasolabial fold area.1 There are no “zero risk” areas on the face, only higher and lower risk areas. This is an important concept that is not well appreciated in the community, and I often hear unsubstantiated claims that such and such technique is safe and effective, since the presenter has been using the technique for years with no ill effect. This is a fundamental error: absence of evidence is confused with evidence of absence. Just because they have not experienced any horrific complication from their technique does not mean that it cannot happen. These are rare events, but because there are large numbers of treatments, serious complications are prevalent (a sort of perverse version of the “law of large numbers”). Let me preface my comments by stating that I really appreciate the efforts these authors have put into doing their due diligence in their thorough literature review and thoughtful analysis of this vexing problem of vision loss following higher risk periorbital facial treatments with HA fillers. My final assessment is that we have not yet reached the final conclusion of this problem. There are several issues and concerns which have not yet been addressed that I hope will serve as a stimulus for further research into this area. The unanswered issues include: (1) Timing of the onset of treatment in relation to the onset of symptoms; can the obstruction be relieved before hypoxia irreversibly damages the relevant tissues? (2) The actual physical location of the arterial obstruction − is the site of arterial obstruction accessible from the retrobulbar space? (3) Is the retrobulbar space similar to other tissue spaces on the face with respect to HYAL treatment? (4)What is the correct dosage of HYAL required to dissolve the obstruction within these blood vessels? and (5) How long does it take to dissolve the obstruction (Table 1)? Factors Relating to Treatment of Filler Arterial Embolism Factors Relating to Treatment of Filler Arterial Embolism The HYAL hypothesis refers to the concept that if an arterial blood vessel is obstructed with HA filler, and the surrounding tissues are flooded with a sufficient quantity of HYAL for a sufficient length of time, the HYAL will diffuse through the vessel wall and break down the HA filler within a period of time (typically a matter of hours).4 There are several assumptions. The vessels must be directly accessible, not within say, a bony canal, or behind an HYAL-impermeable barrier (is the sclera permeable to HYAL?). Even with direct intra-arterial injection of HYAL, the attack surface of HA filler directly accessible to the HYAL is very small, consisting of only the cross-sectional surface area of the blood vessel in question, plus whatever amounts can diffuse through the arterial wall and then back into the vessel further down the obstructed part. In the immersion scenario, the attack surface is much greater, since the HYAL can begin to break down the HA filler in multiple areas as it penetrates the vessel wall. The HYAL itself is being acted upon by dilution, diffusion, and deactivation.5 It is not a static situation as in vitro, since immediately after injection, HYAL is being diluted by local interstitial fluid and serum, diffusing to surrounding tissues, and being deactivated by systemic serum as well as local tissue antihyaluronidases. In addition, we have some indirect clinical evidence to suggest that the periorbital tissues are privileged for HA material, which, you may recall, was originally described in the vitreous humor of the eye by Meyer and Palmer in the 1930s.6 We know from clinical experience that HA fillers typically last from between 6 to 12 months in facial subcutaneous tissues. There is an exception in the periorbital space however. Small amounts of HA filler can cause very long lasting lower eyelid edema for several years (I personally treated several patients with histories exceeding 5 years, far beyond the normal expected duration of effect). When these cases of persistent lower eyelid fullness from displaced HA fillers are treated with HYAL, they typically resolve after one or two treatments. But why didn’t normal HA breakdown occur as with other areas of the face? My pet theory is that there is some antihyaluronidase factor present in the periorbital space that allows for HA fillers to last for extraordinary lengths of time in this space. Evolution may have found it expedient to protect the vitreous humor from breakdown, and thus provides some protection for HA in this space. It may mean that we have to reexamine the idea that the same dose of HYAL will work for treatment of arterial obstruction in the periorbital tissues. As with cardiopulmonary resuscitation (CPR), the timing of the onset of treatment is rather critical, since if treatment is started after necrosis of the affected tissues has begun, then success is impossible. If we take the example of rhesus monkeys in the laboratory, we know that after 97 minutes of clamp time, irreversible changes begin, and that after four hours of complete obstruction, the retinal tissues are irreversibly damaged and nothing we do will restore vision.7 The critical window of treatment onset then is within approximately 97 minutes of vision loss. There really is very little time for referral of the patient to a tertiary care facility in this scenario, and it appears that treatment must be initiated immediately at the site of occurrence if we are to have any possibility of success. However, we have other constraints. We know that even when we have injected HYAL into the tissues in a controlled setting, there is a certain amount of time required for diffusion of the HYAL to the site of obstruction, as well as the time required for HYAL to break down the HA filler into components small enough to pass through the capillary beds. So we have to consider the time required to make the diagnosis, the time to get everything ready, the time to treat (ie, inject the HYAL retrobulbar), the time for the HYAL to diffuse through the vessel walls if they are accessible, and the time for the filler to actually solubilise to the point where the particles are small enough to pass through the capillaries and restore oxygenated blood transport. All of this has to occur within 97 minutes for there to be no permanent retinal damage, and thus it seems improbable at best, even when treatment is provided by the first responder at the bedside (ie, at the injectors office). I’m not claiming that success is impossible, but it does seem to me to be improbable. It’s also clear that we need more information. The other rather obvious problem is that in some regions, the blood vessels are simply not accessible. If the soft tissues around the obstructed vessel are not directly flooded with HYAL, then the HYAL has to diffuse through tissue planes to reach the vessels, or alternatively reach the obstruction by direct intra-arterial injection. Even with super selective angiography, we still have the issue that there is no flow at the site of obstruction, and the HYAL within the vessel lumen has to wash out the blood therein, and then begin to attack the HA exposed to the lumen of the vessel (plus whatever miniscule amount that can diffuse out of the vessel, into the tissues, and then back into the vessel beyond the obstruction). The HYAL will take a relatively long time to work its way through the entire obstruction, and because there is no actual flow, diffusion dynamics determines the rate of advance. The attack surface available to the HYAL is simply the lumen cross-section, plus however far the HYAL can diffuse past the obstruction and back into the vessel lumen − a very small amount of the obstruction. How long does it take to dissolve a thin pipette of HA filler that contains some HYAL at the top? Quite some time, in fact − far longer than the same amount of filler immersed in the same number of HYAL units in my own unpublished little bench studies. Thus, even if we inject sufficient HYAL into the retrobulbar space immediately after visual loss is documented, the HYAL has to diffuse through the retrobulbar space and even then it only has immediate access to the vessels in the immediate vicinity. There is no flow to carry the HYAL to the retina, so it’s back to diffusion dynamics and root mean square timing. It is slow, slow, slow, and the clock is ticking. Meanwhile, the HYAL is likely being deactivated more quickly (my hunch only, no data) than in typical facial subcutaneous tissues, thereby reducing one of the factors (concentration) important in diffusion The concentration of HYAL also appears (in vitro on my own unpublished observations) to also make a significant difference in the rate of filler breakdown. In other words, the same number of international units of HYAL diluted to a larger volume are less effective against a given quantity of filler − dilution slows things down. Serum leaking out of damaged ischemic tissue capillaries dilute the injected HYAL, reducing its effectiveness. The authors quote my bench study from years ago, in which I placed an HA filled, tied off facial artery segment (an “HA sausage”) and bathed it in a solution of HYAL.4 I think there were several limitations in this study. In vitro, there is no dilution, diffusion, or deactivation of HYAL happening as would be expected in vivo.5 The cadaver sourced facial artery segment might have been leaky to HYAL relative to an in vivo specimen. The minimum dose of HYAL required to get an obstruction cleared is not a known quantity, certainly not by my in vitro study. It was a proof of concept bench study that I hoped would be replicated by others in vivo in animal studies, as was done in the excellent work by Kim et al8 and a few others.9 The dose of HYAL to be used in retrobulbar injections is constrained by volume and concentration concerns. We know that very high concentrations of HYAL are possibly toxic to certain cells, at least in tissue culture, and that manufacturers have settled on 150 IU to 200 IU concentrations as safe concentrations for clinical use. Are higher concentrations tolerated by periorbital tissues? Could we use high concentrations of HYAL in an emergency, to save vision? I don’t know. For retrobulbar injections, we also have certain volume limitations which I believe most ophthalmic surgeons believe is about 5 cc for adults, and I am not at all certain when this volume could be repeated without causing concerns about volume of injectate into the limited retrobulbar space. When could we reinject the periorbital space with another dose of 5 mL of HYAL? When does the pressure from repeated retrobulbar HYAL injections start causing perfusion problems? I don’t know. Going back to the clinical observation that HA filler seems to last very much longer in the periorbital space (HA fillers last for several years in the periorbital tissues) raises the possibility that more HYAL may be required in the periorbital space because of (I suspect) anti-HYAL activity in this region. At the very least, we can regard the periorbital space as qualitatively different to the subcutaneous space in other regions of the face with respect to HA duration and possibly to HA dynamics in ways we don’t really understand yet. Therefore, it is likely incorrect to assume that the duration of effect of HYAL in the subcutaneous tissues in rodents (that have muscle tissue − the panniculus carnosis − within the skin) is qualitatively similar to the periorbital/retrobulbar space in humans. My suspicion is that the dosage of HYAL required in the periorbital space may be higher than that required in the subcutaneous space on other parts of the face, but I don’t have any data on this (I admit it is only a clinical hunch). “Here be Dragons” might be a fitting introduction to the filler world, acknowledging the existence of rarely seen complications (some so rare, that one might think that they are only mythical and not a concern). Most individuals will have a happy career injecting dermal fillers, never having to deal with anything worse than a few lumps, or perhaps an occasional asymmetric result. A few unfortunate injectors will indeed “find the dragons.” They’ll find themselves dealing with a serious filler complication, possibly even an irreversible, devastating one such as blindness. Injectors are taught to aspirate before injection but we know that some gels, particularly with the fine gauge needles provided, will test negative despite being right inside an artery.10 There are higher risk areas of the face, particularly the nose, glabella, and nasolabial fold, but there have been reports of blindness from injections in essentially all areas of the face1 − there is no such thing as a completely “risk free” treatment. There are blood vessels everywhere − if it’s alive, it has a circulation, and there will be blood vessels! We can avoid the areas with the named larger vessels, but we should always assume there are going to be vessels wherever we treat. So to my mind, we have higher risk and lower risk treatment behaviors, but there is no such thing as a “risk-free” treatment. Many years ago, Syd Coleman recommended small bolus injections11 as part of a more general avoidance strategy. My unpublished data from the anatomy lab involved a few measurements of the volume of the vessels of the face to the height of the orbit, suggesting that we should keep our bolus size to 0.1 mL or less (in complete agreement with Syd Coleman). More recent measures of specific vessel volumes of the supratrochlear artery, for example, have given us a bit more precision 0.085 mL12 (so the original recommendation was in the ballpark, and probably still a good rule of thumb). Injecting very tiny amounts per pass, especially in the higher risk areas, seems to be one of the most important strategies to me. If, by chance, your needle tip is inside a vessel, keeping the bolus size a small as possible (certainly less than 0.1 mL per pass) reduces the risk of blindness, since that volume shouldn’t reach the retina. I travel to many meetings, and I sometimes watch clinical demonstrations. When I watch injectors squirting in ½ cc or more of filler in one pass, my “spidey senses” go into overdrive. If they only knew about the dragons. This thought-provoking study has demonstrated that late onset intervention is no better than supportive treatment. The authors show that intervention with retrobulbar HYAL is ineffective when “realistic” time lines are followed. Unfortunately, the cases were treated some time after the theoretical ideal treatment window (ie, immediately upon diagnosis). The authors did not prove that retrobulbar HYAL is ineffective in all cases (although I suspect that they are probably correct). It is clear from understanding the timing of the pathophysiology that retrobulbar HYAL injection, if it is to have any chance of working at all, must be done immediately at the filler physician’s clinic as soon as the diagnosis of visual loss is made. Like the brain, the retina is very intolerant of hypoxia. We know that the human retina simply cannot withstand a duration of more than about an hour and a half without any oxygenated blood before suffering permanent injury, and that after four hours, results in catastrophic irreversible loss. This does not mean we should give up. On the contrary, if we can train first responders to do simple CPR and save lives, we should be able to teach injectors safe retrobulbar injection techniques, but we should do so only if controlled clinical trials of immediate retrobulbar HYAL show some benefit for these rare cases. Retrobulbar injection done by the inexperienced will certainly have some associated morbidity, and it remains to be demonstrated that the potential injuries from retrobulbar HYAL are less harmful on balance. Thankfully the numbers of cases are low, in comparison to the number of filler treatments given. Perversely, the rarity of these events are precisely why they are so difficult to study. Regulatory authorities in Canada and the United States have thus far only mandated manufacturers to report AEs. Physicians seem reticent to report these events to colleagues because of the stigma associated with them, and unless regulatory authorities mandate reporting, I don’t see this changing any time soon. Unfortunately, there has not been any significant push to have a complications registry in North America. I personally would welcome a mandatory reporting requirement so that we can not only properly take stock of what is happening now, but also to see the results of any new treatment protocols. Dr DeLorenzi is Medical Director for Allergan Canada (Markham, Ontario, Canada) and Merz Canada (Burlington, Ontario, Canada), and an Advisory Board Member for Kythera Biopharmacueticals, Inc. (Westlake Village, CA, USA), Suneva Medical, Inc. (San Diego, CA, USA), and Valeant Pharmaceuticals (West Laval, Quebec, Canada). The author received no financial support for the research, authorship, and publication of this article.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.