Commentary

In their article, Sutphin et al seek to examine the development of antibodies to a specific formulation of botulinum toxin type A (Botox, Allergan, Irvine, CA) with a seemingly simple and straightforward experiment. The question of antibody formation in response to BTA is obviously of greater import in the functional world of medicine as opposed to the aesthetic world. When treating cervical dystonia or cerebral palsy spasm, doses are much higher, intervals are frequently shorter, and the consequences of losing your best treatment option are much greater than when treating an aesthetic patient. However, loss of the toxin as a cosmetic treatment is obviously something we would rather avoid, however rare those cases might be.

In their experiment, the authors found that the frequency of treatment had a statistically significant effect on antibody production, whereas dose failed to create a statistical significance. They conclude that, “the frequency of administration may represent a more important variable than dosage in terms of eliciting an antibody response.” While that's possible, it may not be entirely accurate for a number of reasons.

Above all, this paper shows that even relatively small doses of BTA elicit antibody production after a single dose. It also shows that this measurable response increases with frequency and seems to increase with dose, although not in a statistically significant manner. Perhaps part of the reason for their breakthrough data is the authors' methodology, because experts in the field of antibody production after exposure to BTA have discussed at length the severe limitations of antibody detection with various existing methodologies.1 Perhaps it is also related to a limitation of the paper to which the authors freely admit: that while they are measuring antibody production, they are not measuring neutralizing antibody production. This lack of specificity does not mean that the antibodies produced were not neutralizing antibodies; it simply means that they non specifically measured total antibody production and, because even the first injections did not provoke a clinical response (presumably because of dose limitations), the percentage of the antibodies that are neutralizing or nonneutralizing remains unknown. The key is that even these subeffective doses clearly stimulate antibody production after a single dose.

The authors make the case that antibody production is a rising problem in aesthetic surgery. In fact, nonresponse or tachyphylaxis is still a very rare problem in the aesthetic world. Data are extremely hard to come by because the event is so rare. Clearly, the consensus among experts in the field is that diminished cosmetic response was virtually eliminated with the BCB2024 BTA batch. In a 2007 article by Borodic1—who publishes more frequently than virtually anyone else on the hazards of antibody production—only a single case report of nonresponse after aesthetic injections is cited. Personally, I know of only a single patient who became a nonresponder after having exclusively been injected with the new BCB2024 BTA batch. This patient worked in another plastic surgeon's office and very frequently was treated with small amounts of “leftover” toxin at the end of the day; after several years of this regimen, she became a nonresponder. When perhaps only thousands of patients were being injected per year, I know of only two treatment failures (in patients that I have seen) from 1991 to 1997. When millions of patients were being injected from 1998 to 2009, I know of only one. I realize how unscientific any general statements made about antigenicity based on the above anecdotal report would be, but it is only anedotal reports of these cases that appear in the literature because they are so rare.

The 20% nonresponder rate cited by Sutphin et al came from a study by Borodic on cervical dystonia patients, who were clearly not receiving treatment for cosmetic use. Even these patients, who received “repeated injections over a period of four to five years,” only “appeared” to have a “neutralizing antibody rate of close to 20%.”2 The italics are mine. Clearly, this is not a hard number from a rigorous study. A recent study of spasticity patients with a mean treatment time of 4.5 years and mean cumulative doses of 4610 U of BTA and 14033 U of abobotulinum toxin A (Dysport; Medicis, Scottsdale, AZ) found antibody presence with clinical relevance (neutralizing antibodies) at a rate of 6%.3 A fascinating study out of the University of Vienna reveals that some dystonia patients have truly neutralizing antibodies (as indicated by smaller anhidrotic areas on the ninhydrin sweat test), but no change in clinical response.4 This is obviously interesting because it calls into question our very definition of neutralizing and nonneutralizing antibodies. Basically, there is more that we do not know than we know about antibodies, toxins, and their associated proteins.

Furthermore, it is always hazardous to extrapolate small animal experiments to humans. Even if we were to assume identical immune systems—which we cannot do—the size of the animal and dose presents a problem, as the authors duly note. The doses used appear to be relatively small, even for a small animal, because no doses elicited a clinical response. If this truly is analogous to small-dose cosmetic injections, the production of antibodies (neutralizing and/or nonneutralizing) may be ominous. However, the rats weighed between 27 and 30 g. Assuming the higher (30-g) weight for the rats and 70 kg for a human, 0.12 U in the rat translates to 280 U in the human, while the higher dose corresponds to 560 U in a human. This is an unfair extrapolation, of course, but when viewed in this fashion, these are not small, cosmetic dose analogs.

The most important thing for a clinician to decide when evaluating the article by Sutphin et al is whether “the frequency of administration may represent a more important variable than dosage in terms of eliciting an antibody response,” as the authors conclude. It is possible, but the problem lies with the variables. The frequencies were planned at one and two months based on the Carruthers chapter5 in Klein's text and an article6 that Carruthers and Carruthers wrote in the dermatologic literature. The authors used their one- and two-month time intervals based on these publications “to simulate a clinically relevant scenario.” However, these are not time limits I would recommend to a cosmetic patient. The authors say that these papers (which are more than 10 years old) advocate no retreatment or boosters within a month of initial treatment. I do not see a rationale for such a recommendation. In my practice, if a patient is noticeably asymmetric, I treat them. I have no proscriptions about the timing. If I retreat the patient shortly after injection (say at one week), I may theoretically be slightly increasing their chances of becoming a nonresponder in the future or slightly hastening the point at which they might become a nonresponder, with the latter probably being more accurate. That must be weighed against the downside of having a dissatisfied, asymmetric patient with untoward results that can be remedied with additional treatment. What I do recommend is that boosters not be given routinely or without justifiable cause, and that patients be made aware of the potential downside of the booster. I know that some experts recommend treating different areas of the upper face two weeks apart to better control final brow position, but I do not personally agree with this.

Although I have obviously listed a series of small quibbles with this article, I remain convinced that this is an important paper with important data. It should be built upon with further study employing more rats, more cohorts at different intervals, and doses that elicit a clinical effect. That way, we will have an idea if these are neutralizing antibodies, nonneutralizing antibodies, or both. The intervals should also mimic clinically relevant intervals in aesthetic injection practices. Widely varying doses would also help to solve another little dilemma, in addition to the booster or touch-up dilemma discussed above.

This is the dilemma of dose as it pertains to axillary hyperhidrosis or extreme sweating. Are we greatly, mildly, or infinitesimally hastening a patient's journey to nonresponse with larger injections to treat his or her axillae, palms, or soles at the same time as a full-face and neck cosmetic injection session? Would separating these doses but shortening the intervals between doses hasten or slow the onset of nonresponse? When we have a series of patients who have been treated for 25 or 30 years, are large percentages going to “hit the wall” and become nonresponders? If they live long enough, will almost all patients or almost none develop neutralizing antibodies? If these patients are then given a long “time out,” will they revert to antibody-negative and responsive status, as was the case with several patients in an older study?7 Currently, I discourage injection into armpits for cosmetic reasons, rather than for true hyperhidrosis, with the explanation that these doses may increase nonresponse. Therefore, I limit my dose and frequency in my cosmetic patients, so that my patients might hopefully enjoy decades of treatment. Am I right? It is the unknown answers to the questions above that make studies like the one carried out by Sutphin et al important for the future of cosmetic medicine. In short, the authors state in their brief conclusion that theirs is a “seminal work,” and I agree with them, but we must analyze these data carefully. The true implications of the work remain to be seen; they are for the readers and time to decide.