Injecting BoNT is a very effective and common noninvasive procedure used to restore facial symmetry and reduce hyperkinesis, synkinesis, and facial imbalance caused by facial palsy [11, 14, 15]. do Nascimento Remigio et al. demonstrated a sustained 9% reduction in facial asymmetry at 6 months after injecting BoNT into the nonparalyzed side of the face [16]. There are numerous guidelines for injecting BoNT into the frontalis muscle based on clinical experiences. Small volumes of BoNT solution are typically injected into from three to nine sites, with doses ranging from 0.5 to 2 U per site. These injections are performed using a 30-gauge, 1-inch needle, which produces a halo of action with a radius of 1.0 ~ 1.5 cm. Progress is often monitored from 1 to 6 months postinjection [8, 14, 16]. The injection technique involves inserting the needle while it is angled 30% upward and extending the injection sites laterally along the lateral canthus [17]. Since the inferior portion of the frontalis muscle intersects with muscle fibers of the procerus and orbicularis oculi muscles, to avoid unexpected complications the recent guidelines suggest that the injections should be performed 2.5–3 cm above the superior margin of the orbit or 1–2 cm above the superior margin of the eyebrow [6, 12, 14, 18]. Another previous study supported that the upper limit of the injection is 1 cm below the hairline, which corresponds to the level of the galea aponeurotica [14]. Despite numerous clinical suggestions, there is a lack of basic anatomical evidence such as about the intramuscular innervation of the frontalis muscle for determining the optimal site of BoNT injection, and there are also no consent guidelines for its implementation in facial palsy [16].
Common side effects of BoNT injections into the frontalis muscle include eyebrow ptosis, an elevated appearance of the eyebrow tail, eyebrow asymmetry, and diplopia [12, 14, 19]. These are technique-dependent complications that may result from the diversity of intramuscular innervation patterns of the frontalis muscle among patients. The appearance of a ‘Mephisto sign’ (‘samurai eyebrow’) can vary with the length, structure, and morphology of the frontalis muscle [19, 20]. To prevent the Mephisto sign, previous authors have recommended performing an additional injection of BoNT into the lateral end of the frontalis muscle, restricted to the area between the midpupillary lines [12]. This recommendation could be explained by the present findings. Because areas A4, B4, and C4 showed relatively high prevalence rates of distal nerve endings (57.9%, 57.9% and 68.4%, respectively), overlooking these areas may result in the Mephisto sign appearing due to compensatory hyperactivity of the lateral end of the frontalis muscle. The shape and size of muscles might also not be the only factors influencing this complication, with the intramuscular innervation pattern (locations of distal nerve endings) possibly also contributing to the Mephisto sign [6, 17].
It has been recommended to inject into the medial part of the frontalis muscle at 1.5–2.0 cm from the facial midline [12]. This is based on the morphology of the frontalis muscle, which has a V-shaped appearance with a slight depression between the bilateral muscles due to midline dehiscence (approximately 3.5 cm above the eyebrow arch). Consequently, administering BoNT at the midline of the forehead may be unnecessary or may require only a minimal amount. This is theoretically supported by our findings of distal nerve endings not being present in area A1 and being present in area B1 in less than half of the specimens. The two sides of the frontalis muscle intersect in area C1, and the angulation between these muscles was more acute in females than in males. Therefore, injecting into area C1 could be expected to be effective in females [21]. On the other hand, injections into area D1 should be restricted to avoid unexpected complications resulting from the frontalis muscle overlapping the procerus, corrugator supercilii, and orbicularis oculi muscles in that area [12].
Previous studies have recommended that the upper limit for injections should be about 1 cm below the hairline due to presence of the galea aponeurotica, with the lower limit being either 2.5–3.0cm from the superior orbital margin or 1.5–2.0 cm above the eyebrow due to the presence of the orbicularis oculi muscle [12, 22]. These clinical recommendations were suggested based on the structural relationships of facial muscles, and they could also be supported by our findings. Distal nerve endings were more prevalent in rows B and C than in rows A and D: there were no distal nerve endings in area A1 in any specimen, and their prevalence rates were low in areas A2, D1, D2, and D4 (31.6–47.4%). Although area D3 showed a relatively high prevalence of distal nerve endings (57.9%), consideration of the structural relationships of the facial muscles may mean that the identified nerve endings were for the orbicularis oculi muscle rather than for the frontalis muscle [20]. Therefore, BoNT injections should not be performed in the uppermost and lowermost areas (rows A and D) due to possibility of side effects and also their low effectiveness.
The central areas in the present study (B2, B3, C2, and C3) are the most-common and well-known injection sites for the frontalis muscle, and presented the highest prevalence rates of distal nerve endings (57.9–89.5%). As references for injection, the midpupillary line is drawn and a site is marked about 2 cm above the superior margin of the eyebrow. Injections are then performed at three sites in a row, spaced approximately 2 cm apart on either side [6, 8, 14, 18]. This 2-cm interval is based on the assumption that BoNT will diffuse approximately 1 cm from the injection site, since the toxin is usually diluted with saline to produce an injected volume of 2.0–3.3 ml. The injection sites of this three-site technique, similar to the central areas in the present study, were chosen because wrinkles are most prominent in these parts of the forehead and they do not overlap with other facial muscles [6, 23]. The findings of the present study have also demonstrated that these injection sites may be most effective.
From a comprehensive consideration of the previously reported information on injection techniques and the results of the present study that utilized objective Sihler’s staining, we propose four injection sites at the boundaries between rows B and C (Fig. 6). Although the dimensions of the forehead differ among individuals, the average forehead height (the vertical distance from the eyebrow to the hairline) and hemiforehead width (the horizontal distance between the facial midline to the frontotemporale) can be considered to be around 5 cm and 6–7 cm, respectively [22, 24, 25, 26]. The present study divided the frontalis muscle on one side into four sets of areas vertically and four sets of rows horizontally. The height and width of each area were approximately 1.25 cm and 1.7 cm, respectively. The commonly injected volume of BoNT can diffuse approximately 1 cm from each injection site (2 cm in diameter), and so administering this at the four sites should result in it covering almost all of the areas with the highest prevalence rates of distal nerve endings whilst also avoiding complications such as the Mephisto sign and eyebrow ptosis. In addition, the physician can easily and rapidly identify these injection sites during the procedure because the distances from the eyebrow to the injection sites are approximately 2.5 cm (1.25 cm + 1.25 cm), allowing them to be measured using two (e.g., the second and third) fingers (Fig. 6).
The arborization pattern of the TBFN refers to the density of the nerve twigs branching off, and this was largest in areas C4 and D4 (Fig. 4). These areas corresponded to where the inferolateral ends of the frontalis muscle were located. This pattern differs from those observed for other skeletal muscles [10, 13, 21, 27, 28]. While the arborization of other skeletal muscles generally forms in the central part of the muscles after where the nerve enters, in the frontalis muscle it occurs at the inferolateral end of the muscle. The TBFN probably exhibits the largest arborization at the inferolateral end of the frontalis muscle due to the topography of the course of the nerve. The TBFN innervates not only the frontalis muscle but also the procerus, corrugator supercilii, and orbicularis oculi muscles. This nerve is therefore already divided into several nerve trunks at its origin from the parotid gland, resulting in the nerve branches becoming thinner, and these fine nerve branches intricately interconnect with each other as they travel toward the frontalis muscle. The abundant branches of the TBFN crossing areas C4 and D4 led to the arborization being largest in these areas. Although the largest arborization and distal nerve endings of the TBFN appeared in areas C4 and D4, BoNT should not be injected into area D4 in order to avoid complications, since this is where the frontalis and orbicularis oculi muscles overlap.
Two types of TBFN arborization patterns were observed. The density of nerve branches in 3 specimens was markedly lower than those in the other 16 specimens (Fig. 5). The variability in the intramuscular innervation pattern of the TBFN could lead to individual differences in the effectiveness of the procedure even when applying the same BoNT injection technique. Physicians should therefore handle cases flexibly according to the symptoms of individual patients based on clear anatomical evidence such as the present research findings.
As the TBFN runs distally to the frontalis muscle it forms fine nerve branches, which results in a methodological limitation of traditional cadaveric dissection for identifying the intramuscular innervation of this muscle. We overcame that limitation by using Sihler’s staining method. The intramuscular Sihler’s neural staining technique is an internationally accepted method for revealing the distribution of intramuscular terminal nerve fiber branches without disrupting the integrity of skeletal muscle. This technique offers the advantage of revealing the relationship between muscle bundles and intramuscular nerve distributions that cannot be observed with the naked eye [29]. The specimen is dyed with 8% Ehrlich’s hematoxylin, which directly highlights the intramuscular terminal nerve branches and their anastomotic relationships. When only the stained nerves are visible, the destaining procedure involves bleaching the muscles and certain surrounding structures while excluding the nerves [10, 29, 30, 31]. Although the neuromuscular junction that absorbs the injected BoNT cannot be observed directly using Sihler’s staining, this technique may still be the most-reliable method for identifying the optimal injection sites for BoNT. Sihler’s staining targets the myelin sheath covering entire nerve fibers, and it was previous reported that the myelinated distal nerve endings are only a few microns from the neuromuscular junction [32, 33, 34]. Therefore, distal myelinated nerve endings as identified using Sihler’s staining can be a reliable marker for the neuromuscular junction [10].
The main limitation of the present study was that we did not directly investigate the position of the frontalis muscle predictable on the facial surface. However, previous studies have already demonstrated palpable surface landmarks on the lateral border of the frontalis muscle. This border runs parallel to a line that extends at 45 degrees from the most-prominent part of the frontotemporal region, positioned 1 cm inward from this line [17, 22, 35]. Based on previous studies it is therefore possible to predict the lateral border of the frontalis muscle using the surface landmark. This allows physicians to determine the 16 areas on a patient’s forehead and use them to identify the optimal 4 injection sites as proposed by the present research.
In conclusion, the present study has objectively verified the scientific relevance of previously reported clinical guidelines for BoNT injections into the forehead and proposed four anatomical injection sites based on objective findings obtained from Sihler’s staining method. However, the present study has also identified that the effectiveness of the procedures could differ among patients even when using the same injection technique due to the variability in the arborization pattern of the TBFN. Moreover, BoNT formulations such as the diluent used (e.g., saline or lidocaine) and the injection volume and technique also vary among physicians [36, 37]. Therefore, it is essential for physicians to consider variables such as symptoms, wrinkle locations, and typical facial expressions of individual patients based on our anatomical suggestions.