A 2012 study evaluated the accuracy of injections into the gastrocnemius muscles of 30 cadavers using palpation and surface landmarks. 121 physicians injected ink into the gastrocnemius muscle followed by dissection to evaluate the accuracy of ink injections (17). Only 43% of the injections were successful with 57% of the injections placed outside of the gastrocnemius muscle, either in the soft tissue (19.8%) or deep to the gastrocnemius, in the adjacent soleus muscle (37.2%). A 2011 blinded study compared "blind" (e.g., manual) versus US placement of a wire into 14 lower limb muscles in fresh cadavers. Two clinicians (a resident with 6 months' EMG experience and an attending > 10 years' experience) performed the needle insertions and the accuracy was then verified by CT, and assessed by a third clinician (18). The overall accuracy in the 14 tested muscles with blind wire placement using anatomic guidance was 39% (range 0%-100%) whereas the accuracy for US-guided wire placement was 96% (range 50-100%). The only muscles in which blind placement was 100% accurate were the tibialis anterior and short head of the biceps femoris, whereas with US guidance the only muscle with less than 100% accuracy was the semitendinosus muscle. Unexpectedly, the accuracy of anatomic guidance/blind placement was 0% for needle insertions into semitendinosus, rectus femoris, and extensor hallucis longus. When comparing the less and more experienced clinicians, there was no significant difference in accuracy of needle placement in the target muscle. The experienced clinician was only more accurate in the trajectory of needle insertion toward the target muscle.
A 2003 study evaluated the accuracy of fine-wire insertions by three physicians (with varying degrees of EMG experience) into 36 lower limb muscles in 10 cadavers (263 muscles) using standard EMG anatomical reference guides (11,19). The accuracy of wire placement was checked by anatomical dissection by an anatomist. In this study, 57% of wire insertions penetrated the target muscle; however, the wire tip was only in the target muscle 45% of the time. As with the previously mentioned 2011 study, there was significant variability in the accuracy of targeting different muscles from 100% for vastus medialis to 0% for 12 attempts in the hip flexors. The proximity of the wire to other structures was also reported with 17% of insertions either penetrating or passing within 5 mm of an important structure. The authors concluded that the accuracy of blind wire placement using EMG reference guides was quite variable and the development of safer strategies was recommended.
Comparison of Anatomic Localization With Other Localization Techniques for BoNT in Limb Muscles
The results from anatomic/cadaver studies are supported by clinical studies evaluating the accuracy of needle placement using anatomic techniques in upper and lower limb muscles.
Manual placement versus EMG: A 2013 prospective study describes the development of a structured protocol using surface anatomy and PROM to localize lower limb muscle injection sites for BoNT injections. The described manual localization protocol details the origin, insertion, innervation, and function of the muscles; how to position the patient, localize the muscle belly, and support the limb, the site, and direction for needle insertions; and the PROM procedure to verify needle location. The authors report that the accuracy of needle insertion using this protocol will be verified using E-Stim. All patients were sedated for the BoNT procedure. While this study provides a description of the protocol, it provides no results from data collection. Therefore the accuracy of this manual/PROM localization protocol is unknown (20).
Manual placement versus EMG: A randomized controlled trial (RCT) in 27 adult patients with spasticity from UMNS (brain injury, spinal cord injury) compared the effectiveness of BoNT injections in upper and lower limb muscles guided by EMG versus injections guided by landmark-based systems. Outcome measures included evaluation of the reduction of spasticity using the Modified Ashworth Scale (MAS) and the functional outcome using the Modified Barthel Index. While a reduction in MAS score and improved Barthel Index score was noted in all subjects, the degree of improvement in spasticity and function was greater in patients where BoNT injections were guided by EMG. The authors concluded that when performing BoNT injections for the treatment of spasticity, the use of EMG to guide injections was superior to injections guided by anatomic/landmark-based reference guides (21). Manual versus EMG: A 2003 study of the efficacy of BoNT injections in adult patients with focal hand dystonia reported superiority of EMG compared to anatomic guidance (22).
Manual placement versus E-Stim: A 2009 study in children with CP (hemiplegia or diplegia) compared the efficacy of BoNT injections guided either by palpation or E-Stim. At 3 months, patients who had injections guided by E-Stim had a statistically greater reduction in MAS scores, PROM, Composite Spasticity Scale scores, and Gross Motor Function Measure scores than those patients injected using manual guidance alone (23). Manual placement checked by E-Stim: A 2005 study in 226 children with CP investigated the accuracy of manual needle for BoNT injections checked with E-Stim to confirm needle position in upper and lower limb muscles (1,376 needle insertions) (24). The needle insertion site was determined by surface anatomy/manual placement and depth estimated by limb size and using PROM. Once the clinician was satisfied with the position of the needle, the accuracy was checked by using E-Stim. During stimulation, the clinician observed the pattern of muscle twitch and whether the twitch response was in the target muscle, or other muscles.
The reported accuracy of manual placement was as follows: gastrocsoleus 78%, hip adductors 67%, medial hamstrings 46%, tibialis posterior 11%, biceps brachii 62%, pronator teres 22%, flexor carpi radialis (FCR) 13%, flexor carpi ulnaris (FCU) 16%, and adductor pollicis 35%. The authors concluded for the muscles tested that manual placement was adequate only in the gastrocnemius. They also postulated that inaccurate muscle targeting could be responsible, at least in part, for a lack of response or insufficient clinical response following BoNT injections in children with CP. Manual placement, E-Stim and US: A 2012 RCT of 49 adult patients evaluated the efficacy of BoNT injections in the gastrocnemius muscle of 49 adult patients with poststroke spasticity (PSS) (25). The study compared the efficacy of three localization techniques (manual, E-Stim, US) using a fixed dose and dilution protocol (onabotulinumtoxinA 100 units medial head, 100 units lateral, dilution 100 units reconstituted with 2 mL preservative free normal saline). At 4 weeks, the patients injected with US guidance had a greater reduction in MAS score than in patients injected with manual needle placement. The US injection group also had a greater increase in PROM compared with the E-Stim and manual injection groups. There was no difference in the Tardieu Scale score between the 3 groups. The authors concluded that when performing BoNT in the gastrocnemius muscle of adult patients with PSS, US guidance provided a greater reduction in MAS and clinical benefit than injections guided with manual needle placement or E-Stim.
Manual placement checked by US: In a 2009 study of 39 children with CP, the accuracy of blind needle placement in the gastrocnemius muscles was checked using US by a blinded clinician (26). The authors reported an overall accuracy of blind placement in the gastrocnemius of 78.7%. The accuracy of needle placement in the thinner lateral gastrocnemius was 64% overall (46% in younger patients), whereas accuracy of needle placement in the medial gastrocnemius was 93% overall (87% in younger patients). The authors concluded that supplementary localization should be considered for the medial gastrocnemius muscle in younger patients and for the lateral gastrocnemius for all patients. Manual placement checked by US: In a second 2009 study of 54 children with CP, the authors evaluated the effect of a number of variables on the effectiveness of lower limb BoNT injections (27). Variables included patient age, BoNT dose, BoNT dilution, muscles injected, and guidance method for injections. BoNT injections were guided by manual needle placement in 44% of patients and by US in 56% of patients. The authors reported a greater efficacy of injections guided by US, in patients less than 6 or greater than 12 years of age, when the muscles injected were hamstrings or gastrocnemii and when the dose/muscle was greater than 0.8 units/kg of onabotulinumtoxinA. Dilution had no effect on efficacy. The authors concluded that this study confirmed the usefulness of US guidance for BoNT injections in lower limb muscles. Location of muscle/muscle fascicles using recommendations from reference guides checked by US: In a 2010 study of forearm flexor muscles in patients with spasticity, the location of the muscle or muscle fascicles using anatomic reference guides was compared to the position of the muscle/muscle fascicle using US. There were significant differences between the estimated position of the muscle or muscle fascicles compared to US for the FCR, flexor policis longus (FPL), and for the fascicles of the flexor digitorum superficialis (FDS) (28).
Although a thorough knowledge of surface and functional anatomy is a requirement when performing BoNT injections and critical for muscle selection, the use of this knowledge alone may not be the optimal method to guide BoNT injections. While many clinicians report a good effect with BoNT using blind needle placement, no study to date has shown this technique to be superior to using a supplemental guidance method. Due to the limitations of anatomic guidance techniques, many if not most clinicians choose to use one or more supplementary localization techniques when performing BoNT injections.