Chan KH, Liang C, Wilson P, Higgins D, Allen GC. Long-term Safety and Efficacy Data on Botulinum Toxin Type AAn Injection for Sialorrhea. JAMA Otolaryngol Head Neck Surg. 2013;139(2):134-138. doi:10.1001/jamaoto.2013.1328
Author Affiliations: Departments of Otolaryngology (Drs Chan and Allen and Messrs Liang and Higgins) and Pediatrics (Dr Wilson), University of Colorado School of Medicine, Children's Hospital Colorado, Aurora.
Objective To evaluate the safety and efficacy data on salivary gland injection botulinum toxin type A for the treatment of sialorrhea.
Design and Setting Retrospective cohort study in a tertiary academic children's hospital.
Patients A 10-year review (January 1, 2001, through December 31, 2010) of 69 children with sialorrhea who had undergone salivary gland injection of botulinum toxin type A.
Interventions Injection of botulinum toxin type A to the submandibular and parotid glands.
Main Outcome Measures Postinjection complications, supplemental treatments, and caregiver satisfaction.
Results A total of 69 children were included in the study (42 boys and 27 girls). The first injection was given at a mean age of 9.9 years with a mean follow-up of 3.1 years. Children underwent ultrasonography-guided 4-gland injection at a constant dosage range. The telephone survey response rate was 51%. Postinjection complications occurred in 19 patients (23 events)—14 (15 events) with minor and 5 (8 events) with major complications. Major complications included aspiration pneumonia (n = 3), severe dysphagia (n = 2), and loss of motor control of the head (n = 3), resulting in 5 hospitalizations and 2 nasogastric tube insertions. Complications were not associated with demographic or clinical factors except for a male preponderance (P = .05). Satisfaction scores were evenly distributed among respondents. Thirty-one children (45%) required supplemental treatments: medical treatment alone (n = 21), surgical treatment alone (n = 2), and combined medical and surgical treatment (n = 8).
Conclusions Although our complication rate is within the published range, some of the major complications had significant morbidity. A subsequent surgical rate of 15% suggests the efficacy is less than universal. Thus, botulinum toxin type A injection for sialorrhea in children is a useful tool but has safety and efficacy limitations.
Botulinum toxin is a protein produced by Clostridium botulinum and is one of the most potent toxins that inhibit neurotransmitter release at the neuromuscular junction. Serologically, there are 7 distinct toxin types, designated A through G. The toxin is a 2-chain polypeptide joined by a disulfide bond. The light chain is a protease that cleaves soluble N-ethylmaleimide–sensitive factor attachment protein receptor (SNARE) proteins.1 Without functional SNARE complexes, the release of acetylcholine from synaptic vesicles into the neuromuscular junctions is inhibited, leading to paralysis of myosin filaments. When neurotransmission at cholinergic junctions in the autonomic nervous system is interrupted, the toxin can also induce various forms of autonomic dysfunction.
The potential clinical use of botulinum toxin type A began in the 1950s when Brooks2 injected the toxin into hyperactive muscles, resulting in temporary reduction of muscle activity. Beginning in the 1960s, Scott3 experimentally injected ocular muscles with botulinum toxin type A in monkeys and subsequently developed treatment for strabismus in humans in the 1970s and received Food and Drug Administration (FDA) approval in 1989. Current FDA-approved indications for the use of botulinum toxin marketed by 5 manufacturers include strabismus, blepharospasm, cervical dystonia, upper limb spasticity, hyperhidrosis, urinary incontinence, migraine headache, and glabellar lines.
The application of botulinum toxin type A for the control of sialorrhea has fascinated physicians for the past 2 decades. It is an off-label use, and to date the mechanism of action at the neuroglandular junction is assumed but not proven.4,5 Its use was first hypothesized by Bushara6 in 1997 for the treatment of sialorrhea in patients with amyotrophic lateral sclerosis. The initial clinical reports of using botulinum toxin type A were focused on adults with neurodegenerative diseases, Parkinson disease,7 and amyotrophic lateral sclerosis.6 The use of botulinum toxin type A in the treatment of sialorrhea in children has only been reported during the last decade. Jongerius et al8 first reported treating 3 children with cerebral palsy (CP) in 2001. The same group published a protocol describing bilateral submaxillary gland injection with botulinum toxin type A under ultrasonography guidance in 2003. Suskind and Tilton9 reported enhanced improvement in sialorrhea control with botulinum toxin type A injection simultaneously into submandibular and parotid glands compared with injecting submandibular glands alone in children with CP in 2002.
The popularity of using intraglandular botulinum toxin type A to control sialorrhea is attributable to the avoidance of invasive surgical procedures but bolstered by its relative excellent published track records of safety and efficacy. Minor complication rates reported in the literature range from 0%9,10 to 11.1%11 for submandibular gland injection and 0%9,10,12,13 to 18.7%14 for submandibular and parotid gland injection. No major complications were reported for submandibular gland injection. Major complication rates ranged from 0%9,10,12,13 to 7.7%14 for combined submandibular and parotid gland injection. The conclusions of these reports suggest that botulinum toxin type A intraglandular injections to the submandibular glands alone or in combination with parotid gland injections for the control of sialorrhea are both safe and efficacious. However, our recent cases with major complications with intraglandular botulinum toxin type A injections warranted a thorough retrospective review of our institutional experiences. Furthermore, it led to a critical review of the published rates of complications to render an opinion on our current practice. These objectives formed the basis for this study.
This study was conducted at Children's Hospital Colorado, a tertiary academic pediatric institution, after approval by our institutional review board. The first part of the study was a 10-year (January 1, 2001, through December 31, 2010) retrospective medical record review of all children younger than 18 years who had undergone intraglandular botulinum toxin type A injection for sialorrhea without prior surgical intervention for sialorrhea. Specifically, information on botulinum toxin type A dosage and complications when present was obtained in addition to pertinent clinical history inclusive of neurodevelopmental and swallowing comorbidities. The second part of the study was a telephone survey conducted on all living children of this cohort obtained from patients, parents, or caregivers. The prospective sialorrhea scoring system was not used at our institution, and a 5-point analog scale (very dissatisfied to very satisfied) was used to assess overall satisfaction.
Complications were classified as minor or major based on outcomes as adopted by the Society of Interventional Radiology.16 Minor complications were defined as requiring no to nominal therapy with no consequence but could include overnight admission for observation. Major complications were defined as minor to major therapy requiring 2 or more days of hospitalization and could result in permanent adverse sequelae or death. Even though this schema was developed for interventional radiologists and their complications are procedure related, this schema was used by Khan et al,14 and we believe it lends itself to standardization of reporting. This schema was used for our cohort and to analyze other publications in the discussion section of this article.
The initial children in this series underwent injection based on dosages and methods described by Suskind and Tilton in 2002.9 These children received 30 U into each submandibular gland and 20 U into each parotid gland. With a desire for increased efficacy and subjective low incidence of untoward effects, the dosages were slowly increased over time. By 2005 a specific clinical protocol at Children's Hospital Colorado for combined submandibular and parotid intraglandular botulinum toxin type A had been developed to standardize the dosing and injection methods. For all patients from 2005 to the present, the following injection procedure was used. The total dose, up to a maximum of 12 U/kg of botulinum toxin type A, was calculated. The final dose used was based on physician preference and previous injection results. Sixty percent of the dose was injected into the submandibular glands (30% per gland), and 40% was injected into the parotid glands (20% per gland). All children underwent injection performed by an interventional radiologist. A pediatric anesthesiologist provided intravenous sedation, using a variety of agents, with cardiopulmonary monitoring in the ultrasonography/interventional radiology suite. Once an acceptable level of sedation had been achieved, one side of the face and neck was prepared and draped using standard sterile technique. The interventional radiologist normalized the concentration of botulinum toxin type A to a concentration of 100 U/mL; therefore, any variation in the volume provided was weight based. Real-time 2-dimensional ultrasonography was used to guide a Chiba-type ultrasonography needle into the center of the submandibular gland, where 30% of the used dose was injected under direct visualization. The needle was removed and redirected under ultrasonography guidance into the same side of the tail of the parotid gland, where 20% of the used dose was injected. The body of the parotid gland was then injected with the remaining 20% without ultrasonography guidance in 3 passes. The same procedure was then completed on the contralateral side. Children were awakened, observed for immediate untoward effects of sedation and/or injection, and then discharged to home.
Beyond descriptive statistics used to summarize patient characteristics, the t test (2-sided) was used to determine an association between complications and age, sex, weight, botulinum toxin type A dosage, and the presence of CP. P < .05 was considered statistically significant.
A total of 69 children (42 boys and 27 girls) with a mean/median age of 9.9/8.9 years and a mean/median follow-up period of 3.1/2.4 years were identified. The demographic characteristics of this cohort are listed in Table 1. It became apparent that the cohort was composed of a skewed population, with 56 of 69 (81%) having swallowing abnormalities and all 69 (100%) having neurodevelopmental diagnoses. Of the 56 children with swallowing abnormalities, 29 had abnormal swallow study results, indicating dysmotility and/or aspiration; 25 were fed by gastrostomy tubes with either normal swallow study results or no swallow studies; and 1 each, as perceived by the physicians or the parents, as having swallowing abnormalities. The neurodevelopmental diagnoses included CP (n = 47), recognizable syndromes (n = 15), chromosomal abnormalities (n = 4), and combined CP and a recognizable syndrome (n = 4). These diagnoses were made by rehabilitation specialists (42%), neurologists (23%), and geneticists (9%).
A total of 120 botulinum toxin type A injections were performed in our cohort during the study period. In total, 23 complications were recorded in 19 patients (16%) during the study period because a patient could have experienced more than one complication (Table 2). Fifteen minor complication events in 14 patients (12%) were identified, of which 6 were related to minor or moderate dysphagia and the remaining had an assortment of minor symptoms. Eight major complication events in 5 patients (4%) were identified, including aspiration pneumonia (n = 3), severe dysphagia (n = 2), and loss of motor control of the head (n = 3). These complications resulted in 5 admissions and 2 nasogastric tube placements. None of the patients required intensive care unit admission; therefore, they were managed with supportive care and intravenous antibiotics during hospitalization.
Comparisons between the complication-free and complication subgroups were performed for the following demographic and clinical variables: age, sex, weight, botulinum toxin type A dosage, the presence of CP, the presence of swallowing disorders, and injection time (before or after 2005). Boys were found to be more likely to develop complications (P = .05). The remaining characteristics were not found to be associated with complications.
A satisfaction survey was completed by 32 of 63 living children in the cohort (51% response rate). The distribution of the respondents was evenly distributed within the 5-point analog scale (Table 3). The combined satisfied and very satisfied groups constituted 44%.
Thirty-one of 69 children (45%) required additional medical and/or surgical treatment in addition to botulinum toxin type A injection (Table 3). Twenty-nine children (42%) received adjunctive medical therapy that included botulinum toxin type B injection (n = 6), transdermal scopolamine (n = 2), and glycopyrrolate (n = 21). Ten children (15%) required sialorrhea operations with (n = 8) and without (n = 2) adjunctive medical therapy subsequent to the failure of botulinum toxin type A to control symptoms. These surgical procedures included rerouting of submandibular ducts (n = 3), ligation of parotid ducts (n = 3), and excision of submandibular glands (n =5).
To our knowledge, this series represents one of the largest cohorts of patients who have undergone simultaneous botulinum toxin type A intraglandular injection of the submandibular and parotid glands for the control of sialorrhea. This series is unique because the cohort had not received prior surgical intervention for sialorrhea as opposed to the series by Khan et al.14 To our knowledge, this series also represents the longest follow-up period after botulinum toxin type A injection. Moreover, the cohort was composed of a population overrepresented by swallowing and neurodevelopmental disorders, which allowed insights into long-term safety and efficacy of this therapy that have not been previously addressed.
The predominance of swallow disorders in our cohort was not limited to those with an abnormal swallow study. Of the 56 patients with swallow disorders, 25 (45%) were fed by gastronomy tube with either no swallow study or normal swallow study results. Although dysmotility and aspiration are hallmark findings of an abnormal swallow study result, other causes and comorbidities may directly and indirectly affect swallowing. Food aversion, nutritional state, and overall neurodevelopmental status are important factors to consider. Thus, gastrostomy tubes were placed in some children who did not undergo swallow studies or in some children with normal swallow study results.
Clearly, the safety component of botulinum toxin type A intraglandular injection, as evaluated by complication rates in the literature, is highly variable among studies. To normalize complication rates across the published series and the present study, the data of our series and in the study by Khan et al14 were converted to reflect the percentage of patient with complication(s) per injection. This normalization was necessary because unlike our cohort and the series of Khan et al, the remaining publications evaluated single rather than repeated botulinum toxin type A injections. Adjusting for complication rate per injection, the reported minor complication rates for 4-gland botulinum toxin type A injection per injection as a whole range from 0%9,10,12,13 to 18.7%14 when the largest series were compared. Our minor complication rate of 12% is within the reported range. The reported major complication rates for 4-gland botulinum toxin type A injection range from 0%9,10,12,13 to 5.5%.14 Our major complication rate of 4% is also within the reported range but remains the second highest rate. Potential factors contributing to these differences include the methods (2 vs 4 glands), botulinum toxin type A dosage, and length of follow-up.
Jongerius et al11 and Scheffer et al,15 who are affiliated with the same institution, injected only the submandibular glands and used a maximum of 50 U of botulinum toxin type A for children who weighed more than 25 kg (maximum dosage, 2 U/kg). Their minor complication rates of 7.7% and 9.2%, respectively, were in the middle of the range as reported in either the 2-gland or 4-gland method. However, neither study encountered any major complications. Both Banerjee et al12 and Reid et al,13 using 2 and 4 U/kg, respectively, reported no minor or major complications using the 4-gland method. Other 4-gland series with smaller sample sizes9 or no follow-up data10 had similar results. This finding would be an argument against the notion that the 2-gland method is safer than the 4-gland method.
Both the report by Khan et al14 and our present study used the 4-gland method with significantly higher mean botulinum toxin type A dosages: 5 and 8 U/kg, respectively. Both reports had the 2 highest major complication rates in the literature: 5.5% and 4.2%, respectively. Although the mean dosage was higher in our study (8 U/kg) than in the study by Khan et al (5 U/kg), a corresponding increase in complication rate was not seen. Curiously, the report by Khan et al and our present report also had the longest follow-up period: 24 and 37 months, respectively. This finding would argue that dosage alone is insufficient to explain these complication rates. Whether repeated botulinum toxin type A injections or lengthening the follow-up period will increase the likelihood of development or observation of minor and major complications is unknown. Thus, we conclude that the increased complication rates for our study and the study by Khan et al do not have an easy explanation.
In addition, the events surrounding the major complications are not trivial. All 5 children who experienced a major complication in this study were hospitalized, and 2 of them required nasogastric tube placement for up to 8 weeks. Given the severity of potential complications, it is prudent for physicians to be aware of these events and ensure that families are well informed. Investigators have postulated that complications might be related to leakage of botulinum toxin type A into surrounding tissue for some time. In 2009, the FDA announced “boxed warnings” on all botulinum products that stated, “botulinum toxin may spread from the area of injection to other areas of the body, causing symptoms similar to those of botulism. Those symptoms include potentially life-threatening swallowing and breathing difficulties and even death.”17 Although the warnings were made in response to the treatment of muscle spasticity in children with CP, patients with sialorrhea share the off-label use of botulinum toxin type A, and thus these warnings may also be applicable to them.
The efficacy of botulinum toxin type A in the treatment of sialorrhea was measured by the different publications using different methods and for various lengths of follow-up. Using standardized although nonvalidated instruments has its advantages. However, these instruments only measure efficacy per the physician or parent/caregiver during a specified period, mostly for 4 to 6 months. In reality, the drooling population requires repeated injections, and a composite picture of the efficacy of therapy may be better assessed from a questionnaire in a retrospective fashion. Across the largest 5 series of botulinum toxin type A studies (regardless of whether parotid glands were injected), the overall efficacy rates had a range of 47% to 83%. Two smaller 4-gland seriesreported even higher efficacy rates of 89%12 and 95%.18 A simple explanation to the diverse efficacy rate is difficult to ascertain, but it appears to favor 4-gland injection. Regardless, this present study clearly points to the need for supplemental medical and even surgical intervention to manage this special population of children with sialorrhea. Yet the overall surgical intervention for sialorrhea at our institution has decreased since using botulinum toxin B as the mainstay medical therapy.
The primary limitation of this study is the retrospective nature of a part of the study. Specifically, a telephone survey is influenced by recall bias, particularly for those individuals/families who had their injections from the earlier part of the study period. Furthermore, the contemporary telephone questionnaire failed to capture satisfaction scores of the entire cohort because of some children's unwillingness to participate or attrition from death due to underlying diseases, which limited our sample size.
The safety and efficacy of a drug are determined by the so-called therapeutic index. The therapeutic index is defined as the “difference between the minimum therapeutic and minimum toxic concentration of a drug.”19 It is also expressed as a therapeutic ratio, which is the dose producing toxicity in 50% of the population divided by the effective dose in 50% of the population. Clearly, none of the studies has achieved the establishment of either a therapeutic index or ratio for botulinum toxin type A. The therapeutic dose of botulinum toxin type A could not be answered by this present study or other published studies. However, on the basis of the results of this study, the clinical course of the children who experienced a major complication, and the dosages used in published series (2 U/kg for 2-gland injections11,15 and <5 U/kg for 4-gland injections14), our institution has decided to half our botulinum toxin type A dose to 4 U/kg for safety and efficacy.
In conclusion, the use of intraglandular injection of botulinum toxin type A in the treatment of drooling has been well established. However, it has limitations on safety and efficacy that can be variable. Patients and parents or caregivers have to be informed of its potential serious complications and the need for adjuvant therapies to achieve the best possible results in controlling drooling.
Correspondence: Kenny H. Chan, MD, Department of Otolaryngology, Children's Hospital Colorado, 13123 E 16th Ave, Ste B455, Aurora, CO 80045 (Kenny.email@example.com).
Submitted for Publication: April 22, 2012; final revision received September 21, 2012; accepted October 30, 2012.
Author Contributions: Drs Chan and Wilson had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Chan, Wilson, Higgins, and Allen. Acquisition of data: Liang, Wilson, and Higgins. Analysis and interpretation of data: Chan, Liang, and Allen. Drafting of the manuscript: Chan, Liang, and Allen. Critical revision of the manuscript for important intellectual content: Chan, Liang, Wilson, Higgins, and Allen. Statistical analysis: Liang. Administrative, technical, and material support: Chan, Liang, Wilson, and Allen. Study supervision: Chan and Allen.
Conflict of Interest Disclosures: None reported.
Previous Presentation: This paper was presented at the American Society of Pediatric Otolaryngology 2012 Annual Meeting; April 21, 2012; San Diego, California and is published after peer review and revision.