New Product Evaluation Basics in Aesthetic Practices

Unique to the field of aesthetics, when a new product such as a neuromodulator is introduced to the market, physicians often elect to gain first-hand personal experience with that product. As they gain confidence in the performance of a product, physicians are able to relay that experience to their patients. However, the method by which a physician chooses to evaluate a new product may have dramatic implications for the interpretation of that product’s performance. Selecting an appropriate single-center study design is critical to controlling for bias that may otherwise be introduced from a plethora of potentially confounding factors. PrabotulinumtoxinA-xvfs (Jeuveau, Evolus, Inc.) is the most recently approved botulinum toxin by the FDA (2019).

There are three basic study designs that a private clinical practice could use to evaluate a new product—each potentially leading to widely different conclusions (Table 1). In each of the three options presented, the group of subjects studied at a single center might include any of the physician(s) themselves, their staff, or some of their patients. In the first option, physicians could simply try out the new product, administering it to all subjects in their chosen test group and observe the results. In this no-control inter-subject design, each subject’s experience would be compared directly with that of the other subjects. Each subject would be treated identically, despite differences in individual needs, and no attempt would be made to account for confounding factors—many of which may be inter-related—that might contribute to a biased result. In the case of a neuromodulator, these sources of bias might include the subject’s age, gender, ethnicity, etc. (see Table 1 footnote). Additionally, should patient-reported outcomes also be collected, subjects would inevitably convey their current experience compared to their recollection of previous treatments with other products.

In the second option (Table 1), one group of subjects would be administered the new product while another group received an active control product. In this active-controlled, inter-subject design, two groups of one or more subjects could be compared. As with the first option, subjects would be treated identically, despite differences in individual needs. However, this type of design, which is typically used in multicenter clinical trials, is not well-suited to a small single center investigation, since large subject numbers and randomization are typically needed to control for the high degree of variability in individual characteristics and demographics that impact the clinical effect experienced by any one subject. Even if randomization was used, this design element must be coupled with large patient numbers to be of benefit in controlling for bias in an inter-subject study.

Importantly, the approach of these first two options is not reflective of current clinical practice. Unlike in clinical trials, patients in an aesthetic clinical practice setting are assessed individually, and neuromodulator treatment is typically customized. However, if implemented in this study design, the number and placement of injection sites, injection depth and technique, and the total dose administered would all play a role in the variability of patient-to-patient outcomes.

In the third option (Table 1), each subject would be administered both the new product and the active control product on opposite sides of their body. Using this type of active-controlled, intra-subject design, only small subject numbers are needed, since each subject serves as their own control, eliminating the subject-to-subject variability seen in the inter-subject design. As such, this format is particularly well-suited to a small private practice clinical setting. Moreover, a higher degree of rigor is introduced where the types of individual subject factors that would otherwise confound the results apply equally to both products. And, more in keeping with current aesthetic clinical practice, it also allows for treatment to be tailored to the individual patient. Accordingly, we undertook a simple N-of-2, split-face study to investigate for ourselves the performance of the prabotulinumtoxinA; onabotulinumtoxinA (Botox Cosmetic, Allergan/AbbVie) was selected as the active control. The clinical effects observed are not only interesting in and of themselves, they also serve to exemplify the importance of study design when undertaking a private clinical practice product evaluation.

AN N-OF-2, SPLIT-FACE STUDY SUITABLE FOR A PRIVATE PRACTICE SETTING

Both “subjects” in our study were adult male 47-year-old, experienced dermatologists who share a clinical practice. Both have their own unique symmetrical bilateral dynamic forehead lines, and thus were each well suited to serve as their own internal control. Each had a unique pattern and severity of dynamic forehead lines at baseline—one relatively mild (Subject 1), the other more severe (Subject 2) (Figures 1a and 2a, respectively). Of the upper facial lines, the frontalis muscle was preferentially selected for this split-face study, since it is a large muscle in which the right and left groups are distinct and well-separated, and any potentially-confounding diffusion of the neuromodulators can readily be minimized by using a midline injection-free zone.

Figure 1a: Subject 1, Relatively Mild Dynamic Forehead Lines at Baseline

Figure 1b: 5 injection sites per side, with prabotulinumtoxinA administered to the patient’s right frontalis and onabotulinumtoxinA administered to the patient’s left frontalis.

Each neuromodulator was reconstituted to a concentration of 2U/0.1mL. The authors injected each other with prabotulinumtoxinA in frontalis sites on one side and onabotulinumtoxinA into frontalis sites on the opposite side; in the case of prabotulinumtoxin A, this is an off-label indication. In order to optimize aesthetic outcomes, the injection protocols were tailored to the particular anatomy and unique pattern of each subject’s dynamic forehead lines. The total dose of each neuromodulator administered was 20U per subject. For Subject 1, that dose was distributed over five frontalis injection sites per side; for Subject 2, that dose was distributed over eight frontalis injection sites per side (Figures 1b and 2b, respectively). The injecting physician was careful to ensure that the target injection sites and the number and depth of injections for their subject were equal left and right. In each case, a minimum midline separation of 2-3cm between the closest target injection sites of the two neuromodulators was maintained.

Figure 2a: Subject 2, More Severe Dynamic Forehead Lines at Baseline

Figure 2b: 8 injection sites per side, with prabotulinumtoxinA administered to the patient’s left frontalis and onabotulinumtoxinA administered to the patient’s right frontalis

Efficacy was followed by video record until the clinical benefit was largely lost in both subjects. Video footage was captured as supplementary digital content (SDC) at Day 0, during the first week post-injection at Day 7, and at Months 3 and 4.

WHAT WE FOUND

For both subjects, early onsets of effect of both neuromodulators were evident by Day 2 post-injection. By Day 7, dynamic lines were no longer in evidence on either frontalis side for Subject 1, and were much reduced for Subject 2. For Subject 1, the clinical effects of both neuromodulators were clearly still in evidence and symmetrical at Month 3. At Month 4, the subject’s dynamic frontalis lines had yet to return to baseline. In contrast, for Subject 2, the clinical effect of both neuromodulators had worn off symmetrically and had begun to dissipate within six weeks of treatment. By Months 3 and 4, little evidence of either neuromodulator remained.

Using this intra-subject, split-face design, we found that 20U prabotulinumtoxinA and 20U onabotulinumtoxinA were similar in their duration in the treatment of dynamic forehead lines. The effects of the neuromodulators began to wear off within six weeks of treatment in one subject and remained efficacious for four months in the other. Clearly, while the two neuromodulators proved to be similar, individual differences in the subjects themselves—in anatomy, physiology, baseline severity, and dynamic muscle activity (as illustrated by these two cases)—had a significant impact on efficacy. This simple design, using an intra-subject comparison and active control (Table 1, design option 3), was key to controlling these individual differences in subject response from what might otherwise have been perceived as differences in the neuromodulators themselves. As such, of the three possible options that could be used by a single center, this study design provides the highest level of reliability when subject numbers are limited.

A different new product evaluation study design would have led to a markedly different conclusion. A no control inter-subject assessment (Table 1, design option 1) would have suggested that the product’s performance was erratic and unpredictable, lasting only six weeks in one subject and four months in the other. Of the three study design options, this has the lowest level of reliability when sample sizes are small. An active-controlled inter-subject comparison (Table 1, design option 2), where one subject received the new product and the other received the control, would have erroneously led to the conclusion that one product had a much longer duration than the other product (i.e., four months versus six weeks). However, the longer duration would have been a function of the product allocation/randomization, rather than the product itself. While randomization could be employed in this design, without large patient numbers, this study design also has a low level of reliability.

Another important feature of our study was that it allowed, in keeping with current guidance, the assessment and optimization of aesthetic outcomes for each subject without confounding the interpretation of results. In recognition of the fact that inter-subject responses to neuromodulators as well as the variations in the frontalis muscle can be significant, the injection protocols were tailored to the particular anatomy, underlying musculature and unique pattern of each subject’s dynamic forehead lines. At the same time, the placement, number and depth of injections, and the total dose were identical for each subject’s left and right frontalis. By controlling for these types of potentially confounding factors through the use of our split-face design, we were able to ensure that the two neuromodulators could be compared directly and on an equal footing.

Another useful aspect of this work is that it can readily be reproduced without the need for the research capabilities of a large center or the more sophisticated metrics and analyses used in clinical trials. By applying common sense and some basic scientific principles that include the use of an appropriate control and study design, a simple evaluation such as this can be conducted within the confines of many clinical practices. This is good news for those who seek out personal experience with a new aesthetic product in order to gain the comfort level that they desire, before recommending a particular treatment to their subjects. As the patient population for these types of aesthetic procedures both grows and diversifies, a split-face study such as this can assist clinicians in gaining the knowledge needed to effectively temper patient expectations with their own real-world experience.

THE BOTTOM LINE

Commonly in aesthetics, clinicians often opt to gain first-hand experience with a new product before recommending it to their patients. How that evaluation is performed is critical. Using a simple intra-subject, split-face design, we were able to demonstrate that prabotulinumtoxinA was similar to an active control, onabotulinumtoxinA even though marked differences were observed between the two subjects studied. This type of experiment, which can readily be conducted even in smaller clinical practices, not only introduced the rigor necessary to control for individual subject factors that would otherwise confound the results, it also allowed for treatment to be tailored to the individual. A less rigorous design could have led to dramatically different conclusions of duration or inconsistent efficacy.

No external funding was provided for this original work. Drs. Hsu and Bhatia have both served as advisors, consultants, and/or investigators for Allergan, Evolus, Merz, and Revance.

Together, Drs Hsu and Bhatia are Co-Directors and Co-Founders of Oak Dermatology, a dermatology and skin care clinic serving the Greater Chicago area.

1. Sundaram H, Signorini M, Liew S, Trindade de Almeida AR, et al. Global Aesthetics Consensus: Botulinum toxin type A – Evidence-based review, emerging concepts, and consensus recommendations for aesthetic use, including updates on complications. Plast Reconstr Surg. 2016;137:518e-529e.

2. Anido J, Arenas D, Arruabarrena C, Dominguez-Gil A, et al. Tailored botulinum toxin type A injections in aesthetic medicine: Consensus panel recommendations for treating the forehead based on individual facial anatomy and muscle tone. Clin Cosmet Investig Dermatol. 2017;10:413-421.