The letter from Dr Frey raised several points worthy of discussion. First was the question of differential species sensitivity to botulinum toxins and whether rabbits are an appropriate model for humans, and specifically whether our rabbits were overdosed relative to typical human usage. Second, the letter noted that we failed to mention the advantages of unloading the temporomandibular joint. Third, it pointed out that our study period of 3 months was not long enough to justify our claim of lasting joint changes. Finally, the letter stated that we claimed that BoNT/A could have direct toxicity for skeletal cells, with no evidence for this claim. We address each issue in turn.

On the question of species sensitivity, we attempted to understand the letter writer’s point by looking up references 3 and 4. Unfortunately, we were unable to locate reference 3 (University of Washington Interlibrary Loan is still looking for a copy of this edited volume), and reference 4 (another edited volume) does not mention botulinum toxin in any of its 741 pages. Nevertheless, species differences do exist, although it is not as simple as herbivores vs carnivores. Mink and foxes are among the carnivores that are highly sensitive to botulism ; pigs, which like humans are omnivores, are much more resistant than humans. There is no evidence that rabbits are more sensitive to botulinum neurotoxin than are humans; both species are subject to the effects of the toxins.

The dose we used (10 units/masseter) falls in the typical range for rabbit studies (5-20 units/masseter) as well as corresponding in dosage to those used clinically. Although the drop-off in electromyographic amplitude was indeed dramatic, the results from human electromyographic studies are similar, not “vastly different.” For example, Kim et al reported 72% to 74% reductions after 2 weeks, actually greater than the 68% reduction we observed at that time point in rabbits. Human masseter electromyographic amplitude continues to be more reduced than that of rabbits at 4 weeks. These human studies were of patients receiving cosmetic reductions of masseter bulk; the patients were not reported to be bruxers or to have abnormal muscle function, and the dosages used (25-35 units/masseter) were in the same range as is used for bruxism. Of course, one should never extrapolate directly from animal studies to the clinic, but we conclude that there is no reason to believe that our rabbits were overdosed relative to clinical usage or that our method was “improper.” Furthermore, a recent rat study and a preliminary clinical study showed that the use of BoNT/A in the masseter is associated with osteopenia of the condyle in 2 additional species. Our result is not unique.

We are not aware of any unequivocal evidence of benefits for unloading the temporomandibular joint and could find no relevant literature. Aside from the obvious point that BoNT/A muscle inactivation and orthodontic appliances are mechanically different and so cannot unload the temporomandibular joint in the “same fashion,” because of the difficulty of measurement in humans it has not been verified that such appliances actually unload the joint. On the positive side, the fact that the condylar cartilage maintained its thickness and proliferative activity suggests that this aspect of joint health was not harmed by unloading. We have no other data to bear on the subject and therefore did not discuss it further. The intent of our contribution was to show that unloading of the joint is most likely associated with loss of condylar bone, a finding that we think practitioners should know about. Clinicians considering treatment with BoNT/A injections into jaw muscles should weigh the potential advantages of unloading against its disadvantages.

We referred to the 3-month duration of effect as “long term” rather than “lasting” because we admittedly do not know whether the effect is permanent. We made no claim that the bone would never recover. As we pointed out in the article, the loss of trabecular connections in the midcondylar region would make it difficult to normalize this region. We agree that longer studies are needed.

With regard to our claim of direct toxicity to bone cells, this is a misreading of our intent. We intended to convey that although we could not eliminate this possibility, the effects observed were more easily explained as a simple consequence of unloading. We do not think that BoNT/A is toxic to the skeleton, but because some organs (eg, salivary glands) do show direct effects of the toxin, we thought we should mention the possibility. However, contrary to the letter’s implication that BoNT/A is good for skeletal cells because it is a treatment for osteoarthritis, we noted that reference 5 of the letter (Chou et al, 2010) is an uncontrolled study that found no evidence for any functional improvement and very marginal evidence for an effect on pain. The moderate analgesic effect of BoNT/A is well known and is not evidence for a therapeutic effect on the skeleton.

In conclusion, we agree with Dr Frey that proper training, case selection, and dosing are vital for clinicians using BoNT/A. There are certainly conditions for which BoNT/A is a treatment of choice, such as dystonias. However, rather than adding to the “myths and misunderstandings surrounding BoNT/A treatments,” we believe that our contribution informs clinicians of a possible side effect that will allow them to improve patient selection.