Thesis

Submandibular duct ligation for drooling in children with neurodisabilities Efficacy, impact and cost-effectiveness Stijn Bekkers

SUBMANDIBULAR DUCT LIGATION FOR DROOLING IN CHILDREN WITH NEURODISABILITIES EFFICACY, IMPACT, AND COST-EFFECTIVENESS Stijn Bekkers

ISBN: 978-94-6421-689-9 Copyright © 2021 Stijn Bekkers All rights reserved. No part of the thesis may be reproduced, stored, or transmitted in any way or by any means without the author’s prior permission, or when applicable, of the publishers of the scientific papers. Cover design: Christine Muris and Nelis Claassen Lay-out: Wendy Bour-van Telgen Printing: Ipskamp The work presented in this thesis was carried out within the Radboud Institute for Health Sciences. The research was financially supported by Johanna Kinderfonds, Arnhem, the Netherlands, Phelps Stichting voor spastici, Bussum, the Netherlands, and Stichting Rotterdams Kinderrevalidatie Fonds Adriaanstichting, Rotterdam, the Netherlands. Printing of this thesis was financially supported by Pentax Medical, Viatris, Oticon, Cochlear, Daleco Pharma, EmiD, Beter Horen, Sonova, ALK, Eurocept, Oticon, DOS Medical, Ipsen, ChipSoft, Soluvos, and Radboud University Medical Center.

Promotor: Dr. F.J.A. van den Hoogen Copromotoren: Dr. A.R.T. Scheffer Dr. C.C.M. van Hulst Manuscriptcommissie: Prof. dr. M.A.A.P Willemsen Prof. dr. A.I. Buizer, Amsterdam UMC Prof. dr. M.A.W. Merkx

SUBMANDIBULAR DUCT LIGATION FOR DROOLING IN CHILDREN WITH NEURODISABILITIES EFFICACY, IMPACT, AND COST-EFFECTIVENESS Proefschrift ter verkrijging van de graad van doctor aan de Radboud Universiteit Nijmegen op gezag van de rector magnificus prof. dr. J.H.J.M. van Krieken, volgens besluit van het college voor promoties in het openbaar te verdedigen op vrijdag 13 mei 2022 om 10:30 uur precies door Stijn Bekkers geboren op 18 juli 1991 te Nijmegen

Supervisor: Dr. F.J.A. van den Hoogen Co-supervisors: Dr. A.R.T. Scheffer Dr. C.C.M. van Hulst Manuscript Committee: Prof. dr. M.A.A.P Willemsen Prof. dr. A.I. Buizer, Amsterdam UMC Prof. dr. M.A.W. Merkx

SUBMANDIBULAR DUCT LIGATION FOR DROOLING IN CHILDREN WITH NEURODISABILITIES EFFICACY, IMPACT, AND COST-EFFECTIVENESS Dissertation to obtain the degree of doctor from Radboud University Nijmegen on the authority of the Rector Magnificus prof. dr. J.H.J.M. van Krieken, according to the decision of the Doctorate Board to be defended in public on Friday, May 13, 2022 at 10:30 am by Stijn Bekkers born on July 18, 1991 in Nijmegen (The Netherlands)

TABLE OF CONTENTS General introduction Chapter 1 Repeated onabotulinum neurotoxin A injections for drooling in children with neurodisability S. Bekkers, T.Y.S. Leow, K van Hulst, L. Orriëns, A.R.T. Scheffer, F.J.A. van den Hoogen Developmental Medicine & Child Neurology 2021; 63 (8): 991-997 Chapter 2 Randomized controlled trial comparing botulinum versus surgery for drooling in neurodisabilities S. Bekkers, C.P.A. Delsing, S.E. Kok, K. van Hulst, C.E. Erasmus, A.R.T. Scheffer, F.J.A. van den Hoogen Neurology 2019; 92 (11): 1195-1204 Chapter 3 Submandibular Duct Ligation After Botulinum Neurotoxin A Treatment of Drooling in Children With Cerebral Palsy S. Bekkers, I.M.J. Pruijn, K. van Hulst, C.P.A. Delsing, C.E. Erasmus, A.R.T. Scheffer, F.J.A. van den Hoogen Developmental Medicine & Child Neurology 2020; 123: 132-137 Chapter 4 An evaluation of predictors for success of 2-duct ligation for drooling in neurodisabilities S. Bekkers, K. van Hulst, C.E. Erasmus, C.P.A Delsing, A.R.T. Scheffer, F.J.A. van den Hoogen Journal of Neurology 2020; 267 (5): 1508-1515 Chapter 5 Surgery vs botulinum for the impact of drooling on daily life in children with neurodevelopmental disabilities: a randomized controlled trial S. Bekkers, I.M.J. Pruijn, K. van Hulst, S.E. Kok, C.P.A Delsing, J.J.W. van der Burg, A.R.T. Scheffer, F.J.A. van den Hoogen DevelopmentalMedicine&ChildNeurology 2021;63(11):1351-1359 11 21 41 63 85 105

Chapter 6 Cost-effectiveness of botulinum neurotoxin A versus surgery for drooling: a randomized clinical trial S. Bekkers, K. van Ulsen, E.M.M. Adang, A.R.T. Scheffer, F.J.A. van den Hoogen DevelopmentalMedicine&ChildNeurology 2020;62(11):1302-1308 Chapter 7 The medium to long-term effects of two-duct ligation for excessive drooling in neurodisabilities, a cross-sectional study S. Bekkers, S. de Bock, K. van Hulst, S.E. Kok, A.R.T. Scheffer, F.J.A. van den Hoogen International Journal of Pediatric Otorhinolaryngology 2021; 150: 110894 Chapter 8 Unsuccessful submandibular duct surgery for anterior drooling: Surgical failure or parotid gland salivation? C.P.A. Delsing, S. Bekkers, K. van Hulst, C.E. Erasmus, F.J.A van den Hoogen International Journal of Pediatric Otorhinolaryngology 2019; 123: 132-137. General discussion Future research - Ethanol submandibular duct ligation for drooling in children with neurodisabilities Summary Nederlandstalige samenvatting Data management form Dankwoord Curriculum Vitae List of publications PhD portfolio 129 149 167 185 197 219 229 241 245 251 255 261

GENERAL INTRODUCTION

12 GENERAL INTRODUCTION GENERAL INTRODUCTION Drooling is defined as the unintentional loss of saliva and other oral contents from the mouth and could be pathological when it persists after the age of 4 years.1, 2 Drooling is a common problem in children with neurodevelopmental disabilities like cerebral palsy (CP). Seventy-eight percent with neurodevelopmental disabilities may have complaints of drooling, with a mean prevalence of 44%.3-7 The prevalence of CP has increased over the past decades and is currently approximately 2 to 2.5 per 1,000 live births worldwide.8-10 But, children with various other neurodisabilities can suffer from drooling as well.11 When treating patients with neurodisabilities, drooling as a problem is easily ignored or overlooked in the wide variety of impairments these individuals are confronted with. Yet, drooling is a disabling condition that leads to both emotional and physical distress: many children consider it their worst condition when it comes to social interaction and relationships.12 The image, odour, and spray of saliva while speaking or coughing can lead to social isolation. Drooling is stigmatizing and leads to intellectual underestimation, which interferes with social integration.13, 14 Drooling patients receive fewer affectionate gestures like attention, hugs, and kisses which can lead to low self-esteem.13, 15-18 Depending on their cognitive and intellectual level, children may be very aware of this.15 Moreover, children who are conscious of drooling interact less with peers.12, 19-21 Drooling does not solely affect the child. Parents or carers suffer practical implications as daily care includes frequently reminding children to swallow pooled saliva, repeated wiping of saliva from lips or chin, and frequent changes of swabs and even clothing throughout the day.22 Physical problems include peri-oral skin erosion, wet clothing, and, when excessive, damage to electronics (including speech computers) or other equipment.23

13 Posterior drooling, saliva that leaks to the oropharynx, might lead to aspiration pneumonia and asphyxia and thereby death. More than 50% of patients with cerebral palsy die from respiratory complications, which is related to aspiration in the majority of patients.24, 25 In conclusion, drooling is associated with physical, psychosocial, and emotional distress, which should be recognized and treated accordingly. Pathophysiology The primary cause of drooling is dysfunctional oral motor control that impairs effective and regular swallowing. Drooling is, however, multifactorial and dependent on various other clinical variables such as gross motor function, dental occlusion, mouth closure, and posture.26-28 Drooling is rarely caused by hypersalivation.29 The submandibular gland is responsible for most salivary production in the resting state (60-70%). The submandibular gland is located in the submandibular triangle, which is bounded by the inferior margin of the mandible, the anterior and posterior belly of the digastric muscle. The gland drains through the 5 cm long submandibular duct (Wharton’s duct), which originates medially to the gland. From here it runs between the mylohyoid and hyoglossus muscle through the genioglossus muscle. The submandibular duct has its opening in the anterior floor of the mouth (sublingual caruncle).30 The sublingual gland, responsible for approximately 10% of saliva production in the resting state, is located just inferior to the floor of the mouth and partly shares drainage with the submandibular gland through the duct of Bartholin, which interconnects the sublingual gland to the submandibular duct. The sublingual gland furthermore drains through approximately 8 to 20 smaller ductules, dubbed the ductuli of Rivini.30, 31 The parotid gland is the largest and is situated pre-auricularly from the zygomatic arch to the angulus of the mandible. The parotid gland drains through the 5cm

14 GENERAL INTRODUCTION long parotid duct (Stenson’s duct) into the mouth opposite to the second upper molar,30 and the parotid gland is responsible for 28% of the saliva in the resting state. When stimulated by chewing, smell, taste or involuntary facial movements, this ratio increases to approximately 53% of the total saliva production.32 Furthermore, there are hundreds of accessory glands in the oral mucosa, which account for approximately 5% of the oral saliva.30 Current treatment and its limitations There is a broad spectrum of treatment options for excessive drooling. Treatments include behavioural therapy, speech and language therapy, pharmacological treatmentwithsystemic anticholinergics, salivary glandsclerotherapy, intraglandular injections with botulinum neurotoxin type A (BoNT-A), and surgery.15, 23, 33, 34 Treatment choice for drooling depends on a wide range of variables such as age, the severity of the problem, type of drooling, oromotor ability, intellectual ability, and intrinsic motivation of patient and caregivers. A stepwise approach for the clinical management of drooling always starts with identifying and optimising conditions that could contribute to saliva loss. Therefore, treatment for drooling in our institute is performed in a multidisciplinary setting that includes a psychologist, speech and language therapist, pediatric neurologist, rehabilitation specialist, and an ear, nose and throat surgeon. A specialized dentist, occupational therapist, and pediatric clinician are consulted upon indication. Although drooling is rarely caused by hypersalivation, several treatments aim to reduce the total amount of saliva, so it suits the child’s oromotor capabilities. Saliva, however, has a wide variety of functions. It is essential in the digestive system because it aids in taste, softens solid foods, and has a function in the breakdown of carbohydrates. Saliva functions as an anti-viral, -bacterial, and -fungal agent, protects teeth, and prevents damage to the oesophagal mucosa by gastroesophageal reflux. Moreover, saliva is important for speaking, the

15 passage of solid foods, and intimate kissing.35, 36 Medication and interventional treatment potentially interfere with the natural function of saliva. There is, therefore, a delicate balance to be found between drooling and xerostomia.37, 38 Behavioural therapy and speech and language therapy are therefore considered first-line therapies for drooling in children with neurodisabilities. Some patients respond and maintain the treatment effect. In other patients, speech and language therapy and behavioral therapy is not effective enough as a single treatment, or it is expected to be ineffective because of the severity of the oromotor disorder or intellectual impairment.39 Anti-cholinergic agents are effective in reducing drooling. However, (systemic) adverse effects are frequently reported and do not uncommonly require cessation of therapy.40-42 Moreover, the effect of oral anticholinergics only lasts as long as the medication is used. The submandibular gland accounts for the majority of the saliva in the resting situation. Most of the invasive treatments in our institution, therefore, primarily target this gland specifically. BoNT-A injections have emerged as first-line pharmacotherapy for drooling resistant to conservative treatment. BoNT-A has proven clinically effective in 50% of children for a median of 22 weeks.15 Adverse events after BoNT-A are usually mild and transient. They primarily include temporary oro-motor paresis resulting in dysphagia. Still, severe dysphagia leading to temporary nasogastric feeding occurs in 3% of patients.43 BoNT-A is injected with ultrasound guidance under general anaesthesia. Due to the inherent temporary effect of BoNT-A, patients generally require at least yearly treatment to maintain the effect.24 Whether it is because the burden of the intervention eventually outweighs the benefits or because the effect of BoNT-A diminishes after repetitive injections, after a certain amount of repeated injections, it is common for patients and caregivers to ask for more definitive treatments. This is when surgery is advocated. In our institution, surgery

16 GENERAL INTRODUCTION is generally considered when the child is older than eight years or when there is a low expectation of “outgrowing” the drooling. Submandibular duct relocation (SMDR) has been the surgical technique of choice for anterior drooling since the 1980s. It is successful in approximately 80% of patients.44-47 SMDR is, in our opinion, still the most effective treatment for drooling, and we believe that this is primarily explained by the relocation of saliva to the base of the tongue, which leads to more frequent swallowing. SMDR is only performed in patients with safe swallowing, adequate posture, and non-progressive neurological disease. When compared to BoNT-A, SMDR is time-consuming and invasive and requires multiple days of admission. Postoperatively we prefer to observe patients one night, intubated and sedated on the pediatric intensive care unit due to the risk of swelling of the floor of the mouth, which can potentially cause upper airway obstruction. Other adverse events include postoperative bleeding, ranulas, nerve damage, sialoadenitis, and transient submandibular gland swelling.48 Briefly, although very effective, SMDR is undoubtedly the more invasive treatment option. Also, SMDR is contraindicated in posterior drooling, unsafe oral intake, and progressive neurodisabilities.48-50 In these cases, submandibular gland excision (SMGE) is commonly advised. SMGE has demonstrated slightly poorer results compared to SMDR but requires just one night postoperative admission. Disadvantages compared to SMDR include risk for adverse events and visible scars near the jawline.51 Severe adverse events after SMDR and SMGE are reported in 8% of our patients and include prolonged intensive care admission and surgical re-intervention.48, 50, 51 The operative time, adverse events, and the necessity for postoperative admission have led us (and others) to consider less invasive alternatives to SMDR and SMGE. A procedure utilizing ligation of the salivary duct, first reported by Klem and Mair in 1999, has slowly gained popularity. Initially, parotid and submandibular duct ligation (four-duct ligation, 4DL) was reported for severe aspiration. Soon

17 after, submandibular 2-duct ligation (2-DL) and 3-duct ligation (ligation of both submandibular ducts and one parotid duct) were reported in the literature. Similar to botulinum toxin injections, 2-DL is a short intervention that takes place in an outpatient setting (Figure 1). The complications after 2-DL seem limited and the success rate of duct ligation in the short term seems only slightly inferior to SMDR and SMGE. Long-term results are, however, ambiguous because recurrence varies between 0 and 58%. Studies yet available are of limited methodological quality and describe relatively small numbers of patients (15 and 12 patients).52, 53 Figure 1. Illustration of the postoperative status after submandibular duct ligation. Duct ligation would appear to have several advantages compared to SMDR and SMGE because it is a significantly shorter surgical procedure (including less extensive dissection) and thus a potentially lower risk for adverse events and complications. However, results in the medium to long-term are not as well established as for SMDR and SMGE. Objectives The primary purpose of this thesis is to study the efficacy of 2-duct ligation (2-DL) to define its position in the palette of interventional treatment for drooling. Recent publications have demonstrated promising results of 2-DL due to its limited invasiveness, limited morbidity, and significant drooling reduction. However, its

18 GENERAL INTRODUCTION efficacy when reducing drooling and improving quality of life has not been studied in a prospective, controlled setting. Similarly, patient selection for the procedure, cost-effectiveness, and the medium to long-term effect has not yet been explored. Outline of the thesis Chapter 1 Botulinum neurotoxin type A (BoNT-A) is currently used as a first-line interventional treatment for drooling. Almost all children exposed to surgery were previously treated with BoNT-A. The effect throughout multiple Botulinum neurotoxin type A (BoNT-A) injections has not been assessed yet. The first chapter aims to evaluate repeated BoNT-A injections to reveal the feasibility of repeated injections in neurodisabilities. Chapter 2 The second chapter primarily compares the effectiveness and adverse events of 2-DL to BoNT-A. BoNT-A served as a control intervention in this randomized clinical trial (RCT) because it is our institute’s first-line interventional treatment for drooling. Chapter 3 Although surgery is effective in the majority of patients, drooling remains refractory to treatment in some patients. The third chapter evaluates the predictive value of submandibular BoNT-A to 2-DL treatment success and compares the effect of 2-DL to BoNT-A within participants. Chapter 4 It has been well-established that surgery for drooling does not always result in favourable outcomes. Drooling is multifactorial, but clinical predictors for treatment success are yet unavailable. Predictors would be useful to guide families and health professionals and would ideally help avoid 2-DL when there is likely to be no positive outcome. The fourth chapter aims to identify predictors for 2-DL treatment success.

19 Chapter 5 Even though BoNT-A and 2-DL lead to a significant drooling reduction, patients are rarely completely dry after either intervention. We question whether we should be satisfied with the treatment effect when there is no improved quality of life accordingly. The fifth chapter aims to evaluate the impact of drooling on daily life and care, social interaction, and self-esteem after either procedure and 2) compare the relief of the impact of drooling on everyday life and care, social interaction, and self-esteem after BoNT-A compared to 2-DL. Chapter 6 Cost-effectiveness analyses (CEA) are becoming increasingly popular as instruments to facilitate health care decision-making. Though CEAs do not appear to reduce costs, they can aid decision-makers in allocating health care resources more efficiently. BoNT-A is currently the most frequently used intervention to reduce severe drooling. There is, however, limited evidence on the costs and cost-relatedeffectiveness of BoNT-A for the treatment of drooling. 2-DL might rival BoNT-A because of the effect and little related morbidity. The sixth chapter aims to study the costs and cost-related effectiveness of 2-DL compared to BoNT-A after one treatment cycle. Chapter 7 There is little research available that studies the long-term effect of 2-DL. The seventh chapter aims to investigate the subjective 2-DL anterior drooling reduction in the medium to long-term. Chapter 8 There is some degree of recurrence after submandibular duct surgery and in some patients drooling seems refractory to interventions in general. The eighth chapter primarily aims to evaluate the effect of revision surgery for drooling, and secondarily seeks to identify patients characteristics associated with failure of submandibular duct surgery.

1 REPEATED ONABOTULINUM NEUROTOXIN A INJECTIONS FOR DROOLING IN CHILDREN WITH NEURODISABILITY Stijn Bekkers, Theresa Y.S. Leow, Karen van Hulst, Lynn B. Orriëns, Arthur R.T. Scheffer, Frank J.A. van den Hoogen Published in: Developmental Medicine & Child Neurology 2021; DOI: 10.1111/dmcn.14872.

22 CHAPTER 1 ABSTRACT Aim To evaluate the effect of repeated onabotulinum neurotoxin A injections for the treatment of drooling in children with neurodisabilities. Methods This was a retrospective cohort study, in which the first, second, and third onabotulinum neurotoxin A injection were compared within children treated between 2000 and 2020. Primary outcomes included drooling quotient, visual analogue scale (VAS), and treatment success defined as ≥50% reduction in drooling quotient and/or VAS 8 weeks after treatment. Each outcome was obtained at baseline and 8 weeks posttreatment. Results Seventy-seven children were included (mean age at first injection: 8y 3mo, SD 3y 7mo, range 3 – 17y; 44 males, 33 females; 51.9% with cerebral palsy, 45.5% wheelchair-bound). The objective (drooling quotient) and subjective (VAS) effect after the second injection was lower compared to the first injection. The third injection showed less objective and significantly less subjective effect compared to the first injection. An overall success rate of 74.0%, 41.6%, and 45.8% were found for the first, second and third injection respectively. Interpretation Although onabotulinum neurotoxin A remained effective throughout the entire treatment course, there is less effect of subsequent onabotulinum neurotoxin A injections compared to the first. Although there might be a loss of effect after repeated injections, there is continued improvement in a vast degree of the children. What this paper adds • Repeated injections show a diminished treatment effect after the second injection. • A continued improvement is seen in most patients.

23 Repeated onabotulinum neurotoxin A injections for drooling in children with neurodisability 1 INTRODUCTION Drooling, defined as the unintentional loss of saliva, is considered atypical when it persists after the age of 4 years.1 Children with cerebral palsy or other neurodisabilities often suffer from drooling.2 Drooling interferes with social interaction, and leads to intellectual underestimation and low self-esteem.3 When treating children with neurodevelopmental disabilities, drooling is easily overlooked. However, patients often consider it as one of their worst affections in relation to social interaction.3 Onabotulinum neurotoxin A (Botox; Allergan, Nieuwegein, the Netherlands) is currently the first-line interventional treatment for drooling because it is effective in the majority of patients, minimally invasive, and there is limited risk for severe adverse events.4 Yet, the effect is temporary, and to maintain effect injections are required at least once per year.4,5 Moreover, onabotulinum neurotoxin A injections generally take place under general anaesthesia and one injection necessitates multiple hospital visits including an anaesthesiologist visit and to monitor the effect on drooling.6,7 These drawbacks may lead to discontinuation of treatment. Another possible reason for discontinuation is the diminished or lack of effect after repeated injections.8 Neutralizing antibodies (NAb) or parotid gland compensation may play a role in the decrease of the effect after repeated injections, while alternative literature hypothesize gland atrophy and subsequently a permanent reduction in drooling.9–13 The aim of this study is to evaluate the effect of repeated onabotulinum neurotoxin A injections to reveal the usefulness of repeated injections for the treatment of drooling in neurodisabilities.

24 CHAPTER 1 METHODS A retrospectiveobservational studywas conducted in childrenwhowere treatedwith at least two subsequent, equally dosed onabotulinum neurotoxin A injections. The effect of onabotulinum neurotoxin A injections was compared within participants to reduce confounding factors and increase reliability in a heterogeneous population. Study population Children aged 4 years or older treated for drooling at the Radboud University Medical Center, Nijmegen, the Netherlands, between 2000 and 2020 were eligible. Children treated for anterior or antero-posterior drooling with at least two onabotulinum neurotoxin A injections with identical dose (25IU/gland) delivered at the same gland(s) were included in this study. The first injection in this study was the first onabotulinum neurotoxin A injection that was administered for the treatment of drooling. Both injections in the submandibular glands and submandibular and parotid glands (combined injection) were included. In our centre the parotid glands are not routinely treated because the submandibular glands are thought to be responsible for 70% of the saliva in a resting situation.14 Children were excluded if dosages of the first or second injection were unknown. Assessments of subsequent injections different in onabotulinum neurotoxin A dose or gland localization were excluded, but previous equal injection within the same children were included. Also, children with a progressive condition (i.e. mitochondrial diseases and metabolic diseases) were excluded. The data were saved in an anonymized protected web-based database according to good clinical practice. This study was approved by the medical ethical committee of the Radboud University Nijmegen Medical Centre (CMO: 2020-6145). Outcome measures Regular standardized assessments were made by a specialized speech and language therapist and included the severity of drooling using the drooling quotient

25 Repeated onabotulinum neurotoxin A injections for drooling in children with neurodisability 1 and a visual analogue scale (VAS) before the injection (baseline) and 8 weeks after the injection. The most recent baseline measurement was used in case of multiple assessments. Primary outcomes The drooling quotient and VAS at the 8-week assessment were used as primary outcome variables. Reduction was defined by the baseline minus the outcome at the 8-week assessment. The assessment at 8 weeks was chosen as onabotulinum neurotoxin A injections have shown a maximum effect of clinical reduction in drooling between 2 and 8 weeks.6 A clinically meaningful change (treatment success) was defined as ≥50% reduction in drooling quotient and/or VAS 8 weeks after treatment.4,5,15,16 Treatment success was calculated by either VAS or drooling quotient alone in case of a missing drooling quotient or VAS. The drooling quotient is a direct, observational, semi-quantitative method validated to evaluate anterior drooling.17 It is used as measurement for the severity of drooling by observing actual salivation over a 5- or 10-minute period of time during activity or rest. Both are equally validated procedures, but assessment during activity has been shown to distinguish drooling severity better than during rest.17 In this study, the drooling quotient during activity was used. The type of activity varied from playing with blocks and using electronic devices, and was adjusted based on the child’s abilities and interests. The 5- and 10-minute versions were both accepted as they can be used interchangeably.17 Subjective drooling is assessed by a VAS for severity of drooling. Parents or caregivers are asked to indicate drooling severity on a line ranging from 0 (no drooling) to 100 (being the most severe outcome) for the severity of drooling over the past 2 weeks.18 Secondary outcomes Adverse events documented are described. In addition, the drooling severity and drooling frequency scale are presented in Table S1 (online supporting information).

26 CHAPTER 1 Statistical analysis All statistical analyses were performed using IBM SPSS Statistics 25 (IBM Corp., Armonk, NY, USA). For each outcome measure (drooling quotient or VAS at the 8-week assessment), a multilevel linear model was used to account for the relation between the number of repeated injections (ordinal variable) within individuals. Sample size estimation was not suitable because of the retrospective nature of the study. This model included fixed effects for baseline (drooling quotient or VAS respectively), age, sex, cerebral palsy, degree of mobility (defined by ambulant or non-ambulant), developmental age (defined as <4y or >4y), and epilepsy and random effects for intercept and participants. Predictors were deemed as fixed as they generally did not change through time. There was no need for imputation of missing data in the outcome variables as multilevel linear models are competent in handling these missing values.19 A paired sample t-test and repeated analysis of variance (ANOVA) was used to analyse the differences in baselines of drooling quotient and VAS and the differences in the 8-week assessments between consecutive injections. To evaluate potential bias due to a difference in duration between baseline and its respective follow-up assessment (as the 8-wk assessment does not always take place at exactly 8wks after injection), a paired sample t-test was performed. A paired sample t-test was also performed to evaluate differences between injections in the duration between the onabotulinum neurotoxin A injection and the subsequent baseline measurement, because the previous injection may influence the next baseline through ongoing effect. RESULTS Of all the children treated with onabotulinum neurotoxin A injections for drooling between 2000 and 2020, 85 were treated with at least two subsequent, identical onabotulinum neurotoxin A injections. One patient turned 4 years old 3 weeks

27 Repeated onabotulinum neurotoxin A injections for drooling in children with neurodisability 1 after the first injection and was therefore included for analysis. Nine children were excluded because of a progressive condition or unknown first or second dosage of onabotulinum A injection. In total, 77 children were included in the analysis (mean age 8y 3mo, SD 3y 7mo, range 3-17y; 44 males, 33 females). Twenty-four children (31.2%) had an identical third onabotulinum neurotoxin A injection as well (Table 1). Out of the 356 drooling quotient and VAS measurements, there were 29 (8.1%) missing values for drooling quotient and 29 (8.1%) missing values for VAS. The mean drooling quotient and VAS scores per assessment are shown in Figure 1. Table 1: Baseline patient demographics Patient characteristic n (%) Age, mean, SD (range), y:mo 8:3, 3:7 (3–17) Sex Male 44 (57.1) Female 33 (42.9) Main diagnosis Cerebral palsy 40 (51.9) Non-cerebral palsya 37 (48.1) Degree of disabilityb Ambulant 42 (54.5) Non-ambulant 35 (45.5) Developmental age <4y 49 (63.6) >4y 28 (36.4) Epilepsy Diagnosed with epilepsy 42 (54.5) No epilepsy 35 (45.5) Type of injection Submandibular 72 (93.5) Submandibular and parotid 5 (6.5) aNon-cerebral palsy consisted mainly of developmental disorders based on a syndrome, metabolic, or genetic disorder. bThe degree of disability in children with cerebral palsy was based on the Gross Motor Function Classification System (GMFCS). A GMFCS level of I–III was classified as ambulant and a GMFCS level of IV–V was classified as non-ambulant.

28 CHAPTER 1 Figure 1: Mean drooling quotient and visual analogue scale (VAS) over time. Linear graph showing the trend in time of the mean drooling quotient and VAS at baseline and 8-week assessment per injection. Error bars show standard error. Our analysis showed that the number of injection (injection 1, 2, or 3) was associated with a lower value indicating improvement on the drooling quotient (p=0.016) and VAS (p=0.002) after 8 weeks. This indicates that the subjective (VAS) and objective (drooling quotient) reduction is lower after the second injection than after the first injection (slope [b]=–5.20; 95% confidence interval [CI]=–8.72 to –1.68 and b=–12.74; 95% CI=–19.78 to –5.70 for drooling quotient and VAS respectively). The reduction (∆) in VAS was lower after the third injection compared to the first injection (b=–10.35; 95% CI=–20.62 to –0.87) (Fig. 2). The analysis was corrected for age: 8 years 2 months for (∆) drooling quotient and 8 years for (∆) VAS. The reduction in drooling quotient did not differ significantly between the third and first injection.

29 Repeated onabotulinum neurotoxin A injections for drooling in children with neurodisability 1 Figure 2: Stacked diagram showing the trend of predicted drooling quotient (DQ) and visual analogue scale (VAS) per injection. Baseline is represented as the predicted reduction and predicted outcome combined. Predicted (adjusted) baseline values across all three injections were 26.7 and 74.3 for drooling quotient and VAS respectively. Level of significance: ap<0.05; bp<0.005 Significant predictors for both drooling quotient and VAS were baseline and age. Namely, a higher baseline drooling quotient or VAS predicted a higher outcome (drooling quotient and VAS) after 8 weeks, as well as a greater reduction in drooling quotient and VAS, and a higher age was associated with less reduction in both drooling quotient and VAS. The degree of motor disability predicted VAS; nonambulant children (defined by a Gross Motor Function Classification System level of IV or higher) have a greater reduction in VAS than ambulant children. No association was found for sex, developmental age, cerebral palsy, and epilepsy. The 8-week assessment of the drooling quotient and VAS after the second injection and the VAS after the third injection were slightly but significantly higher than after the first injection (Fig. 1). The second and third VAS baselines were significantly lower than the first VAS baseline. In case of drooling quotient, only the second

30 CHAPTER 1 baseline was significantly lower than the first baseline. The third drooling quotient baseline was not significantly different compared to baseline 1. An overall clinically meaningful change was observed in 74.0%, 41.6%, and 45.8% after the first, second, and third injection respectively. A paired t-test showed no significant time difference in duration of assessment between the injections. The mean duration between the first injection and the subsequent baseline was 50.3 weeks (SD 40.8). The mean duration between the second injection and third baseline was 61.6 weeks (SD 52.4). A repeated ANOVA with a Greenhouse–Geisser correction showed no significant difference of the duration between the day of injection to the actually performed 8-week follow-up per each injection (first injection 9.1wks, SD 2.0 wks, second injection 9.0wks, SD 1.5wks, third injection 9.5wks, SD 3.8wk). Potential complications and adverse events Adverse events after each onabotulinum neurotoxin A injection are reported in Table 2. No serious adverse events were reported. χ2 tests show no significant differences between injections in the overall frequency of adverse events. Table 2: Adverse events and complications after onabotulinum neurotoxin A injections reported by parents Injection #1 #2 #3 Adverse events, n (%) 12 (15.6) 10 (13.0) 4 (16.7) Xerostomiaa 5 3 1 Diminished feeding or discomfort during feeding 4 1 2 Diminished drinking 0 1 1 Saliva swallowing problemsb 6 7 2 Pneumonia (possibly due to aspiration) 0 1 0 Teeth grinding 0 0 1 Bruises injection location 1 0 0 aXerostomia included dry mouth and dry lips. bProblems with saliva swallowing included changes in viscosity, increased choking/gagging, and discomfort during swallowing.29

31 Repeated onabotulinum neurotoxin A injections for drooling in children with neurodisability 1 DISCUSSION This study evaluated the effect of repeated onabotulinum neurotoxin A injections in children with drooling. Although there was continued improvement after repeated injections in the vast majority of children, this study also revealed a significant (but slight) objective (drooling quotient) and subjective (VAS) decrease in effect when comparing the first injection to the second and third injection. Moreover, treatment success rates seemed to decline between the first and second, but not between the second and third injection. Each injection showed similar adverse events and no serious adverse events, supporting the safe use of onabotulinum neurotoxin A, also after repeated injections. There are several explanations for the decrease in effect after repeated injections. The second and third baselines for both VAS and drooling quotient were lower than the first baseline. This could be due to an ongoing treatment effect at the baseline measurement of the subsequent injection, gland atrophy, or children who mature and outgrow drooling. However, baselines before each injection have been corrected for in the analysis. Moreover, although one could expect a lower 8-week outcome after the second injection due to: (1) a lower baseline, (2) survivor bias, or (3) presumably gland atrophy, both 8-week drooling quotient and VAS were slightly higher at the second injection compared to the first injection. Therefore, the decrease in effect cannot be neglected. Another explanation for the decrease in effect could be parotid compensational hypersalivation. One study mentioned dilated ducts and more mucus accumulation in parotid glands after resection of the submandibular glands in rabbits.20 Another study revealed increased parotid salivary flow rate after botulinum neurotoxin A injections in the submandibular glands, which indicated parotid gland compensation.13 Correspondingly, most children in our study received onabotulinum neurotoxin A injections in the submandibular glands only, leaving the parotid glands untreated, which could have led to compensational hypersalivation of the alternative glands. This would, however, have resulted in higher subsequent

32 CHAPTER 1 baseline outcomes which it did not. Alternatively, NAb against botulinum neurotoxin A have been suggested to play a role in secondary treatment failure, mostly injected in muscles to decrease spasticity.10–12 One case report described secondary non-response of botulinum neurotoxin B injections in drooling after three successful injections. NAb against neurobotulinum toxin B were identified and additional onabotulinum neurotoxin A injection did not show treatment response either.21 Some children included in this study may also have been treated with botulinum neurotoxin A injections in their limbs because of spasticity which may have induced NAb before the first intraglandular onabotulinum neurotoxin A injection. In a recent cross-sectional study, the prevalence of NAb against onabotulinum neurotoxin A in children treated for spasticity was found to be 15.2%. Also, the study found the most influencing factor to be the single dose per session.9 As this study was done retrospectively, data on whether children were treated with botulinum neurotoxin injection for other indications could not be assessed thoroughly. Although this study certainly revealed a significant reduction in effect between the second and first injection, there was no ongoing decrease in reduction after the second injection while NAb development during onabotulinum neurotoxin A treatment most likely increases with treatment duration.9 Yet, the study was subject to survivor bias, so children with loss of treatment effect after the second injection will presumably not undergo a third injection, indicating that only responders are included for subsequent injections. In summary, there was presumably an ongoing effect of onabotulinum neurotoxin A on drooling which resulted in a lower second and third baseline outcome. Though there was no ongoing decrease in objective effect after the third injection, the decrease in effect compared to the first injections is not neglectable. Future research should evaluate whether NAb play a role in the decrease in effect after repeated botulinum neurotoxin injections in a prospective setting. Despite the decrease in effect after repeated injections, onabotulinum neurotoxin A still

33 Repeated onabotulinum neurotoxin A injections for drooling in children with neurodisability 1 improved drooling in our children during the entire treatment course (Fig. 1). In case of lack of response after repeated injections, combined injections (submandibular and parotid) could be considered. For children aged above 10 years or with low expectation to ‘outgrow’ drooling, we consider surgery. Unlike the objective effect, the subjective effect of the third injection decreased as well. We interpret that the lack of objective effect could be due to the small sample size, parental expectations, or the burden on carers which might have outweighed the benefit of the procedure. Patients may have forgotten the severity of drooling before treatment, and, as such, have a high VAS score once the drooling has resurfaced again. Studies which evaluated the effect of repeated botulinum neurotoxin A in children are scarce.22 One prospective study evaluated the effect of repeated botulinum neurotoxin A andB in parotid and submandibular glands for drooling in childrenwith neurodegenerative disorders. Five out of 30 children showed no effect (decrease in Teacher Drooling Scale) on repeated injections, but it was not mentioned at which injection, so we cannot draw definitive conclusions from this article.22 Another study identified a significant decrease in effect after repeated botulinum neurotoxin B injections for the treatment of drooling, but only after the fifth injection. This study, however, only included a small population of 17 children.23 Furthermore, studies on the effect of repeated botulinum neurotoxin A injections for other indications (e.g. spasticity, epiphora, and axillary hyperhidrosis) also showed a consistent effect.24–27 One prospective observational study found a diminished effect after two to three injections in lower extremity spasticity in children with cerebral palsy.28 Nonetheless, botulinum neurotoxin A was given for other indications and this finding is therefore not fully applicable to our aim. Still, it does show the decrease of effect after repeated use of botulinum neurotoxin A in children. This study also found a treatment success of 74%, defined as ≥50% reduction in drooling quotient and/or VAS at 8 weeks compared to baseline, in this population

34 CHAPTER 1 for the first injection. This is relatively high compared to: (1) treatment success (53–70%) of previous studies at the same centre using similar definitions and (2) treatment success of the second and third injections in the current cohort.4,5,29,30 Two studies, however, evaluated the effect of onabotulinumneurotoxin A in children who received an average of 1.6 to 2.3 previous injections whereas our success rate is calculated solely based on the first injection.4,5 As our inclusion criteria only allowed children with repeated injections, survivor bias may have caused a higher treatment success rate for the first injection. In addition, non-ambulant children show better subjective effect compared to ambulant children, but were treated with fewer injections. Reasons for discontinuation of treatment, despite satisfactory response rates, could include the practical burden of repeated hospital visits for wheelchair-bound children. Age was negatively correlated with treatment effect: older children respond less. This is in contrast to a recent study in which age (more mature) was considered a predictor for successful two-duct ligation treatment of anterior drooling.15 We cannot fully explain this, but one theory is that older children hadmore exposition to previous onabotulinum neurotoxin A which could have induced NAb. Alternatively, expectations of carers may be higher after repeated injections. However, this may not be the only reason as the objective measure is also negatively correlated with age. The main strengths of the study were the number of children included, the standardized subjective and objective measurements, and the homogeneity of treatment characteristics. There are, however, also some limitations to the study. First, as this study was done retrospectively, treatment was clinically driven. Children with a good response to onabotulinum neurotoxin A will most likely continue with the same treatment, potentially leading to survivor bias. This may result in an overestimation of the effect of the second and third injection. Second, there was a relatively small number of children included with three injections, which means that the study had relatively little statistical power to address the effect of the

35 Repeated onabotulinum neurotoxin A injections for drooling in children with neurodisability 1 third injection. Lastly, our definition of clinical success is defined by at least 50% decrease in the drooling quotient or VAS, which may result in limitations because of its dichotomous nature.30 Conclusion Onabotulinum neurotoxin A remained effective throughout the entire treatment course and induced lower baseline levels as well. However, this study reveals a reduced subjective and objective effect of subsequent injections compared to the first onabotulinum neurotoxin A injection and, as such, there is possibly a limit to the effect of repeated onabotulinum neurotoxin A for the treatment of drooling in children with neurodisabilities. This reduction may be (partially) explained by NAb or compensational salivation by alternative glands. Importantly, although there might be a loss of effect after repeated injections, there is continued improvement in a vast degree of the children. Future prospective research should further evaluate the precise clinical role of NAb and compensational salivation after repeated onabotulinum neurotoxin A injections for the treatment of drooling. Acknowledgements Peter Jongerius performed most of the botulinum neurotoxin injections. We want to thank Corrie E Erasmus, Marloes Lagarde, and Sandra de Groot for their contribution in patient care. This study and its analysis were subsidized by Johanna Kinderfonds, Arnhem, the Netherlands, Phelps Stichting voor spastici, Bussum, the Netherlands, and Stichting Rotterdams Kinderrevalidatie Fonds Adriaanstichting, Rotterdam, the Netherlands. All authors report no conflict of interest. None of the authors reported financial disclosures. Data availability statement The protocol, anonymized demographics, and data regarding study outcomes will be shared on request from any qualified investigator.

36 CHAPTER 1 Supporting information The following additional material may be found online: Table S1: Drooling severity and drooling frequency Drooling Severity No (1) Mild (2) Moderate (3) Severe (4) Profuse (5) Missing #1 0 (0%) 0 (0%) 0 (0%) 11 (14.3%) 52 (67.5%) 14 (18.2%) #1 (8weeks) 0 (0%) 7 (9.1%) 14 (18.2%) 19 (24.7%) 23 (29.9%) 14 (18.2%) #2 0 (0%) 1 (1.3%) 8 (10.4%) 17 (22.1%) 48 (62.3%) 3 (3.9%) #2 (8weeks) 0 (0%) 2 (2.6%) 17 (22.1%) 24 (31.2%) 26 (33.8%) 8 (10.4%) #3 0 (0%) 0 (0%) 1 (4.2%) 8 (33.3%) 14 (58.3%) 1 (4.2%) #3 (8weeks) 1 (4.2%) 1 (4.2%) 8 (33.3%) 3 (12.5%) 10 (13.0%) 1 (4.2%) Drooling Frequency Never (1) Occasional (2) Frequent (3) Constant (4) Missing #1 0 (0%) 2 (2.6%) 21 (27.3%) 40 (51.9%) 14 (18.2%) #1 (8weeks) 0 (0%) 37 (48.1%) 24 (31.2%) 2 (2.6%) 14 (18.2%) #2 0 (0%) 11 (14.3%) 36 (46.8%) 27 (35.1%) 3 (3.9%) #2 (8weeks) 1 (1.3%) 29 (37.7%) 33 (42.9%) 6 (7.8%) 8 (10.4%) #3 0 (0%) 3 (12.5%) 17 (70.8%) 3 (12.5%) 1 (4.2%) #3 (8weeks) 2 (8.3%) 13 (54.2%) 2 (8.3%) 6 (25%) 1 (4.2%) Percentages per assessment (baseline/8weeks) per outcome. 77 patients included for #1 and #2. #3 consists of 24 patients.

37 Randomized controlled trial comparing botulinum versus surgery for drooling in neurodisabilities 2 REFERENCES 1. Blasco PA, Allaire JH. Drooling in the developmentally disabled: management practices and recommendations. Consortium on Drooling. Dev Med Child Neurol 1992; 34: 849–62. 2. Speyer R, Cordier R, Kim JH, Cocks N, Michou E, Wilkes-Gillan S. Prevalence of drooling, swallowing, and feeding problems in cerebral palsy across the lifespan: a systematic review and meta-analyses. Dev Med Child Neurol 2019; 61: 1249–58. 3. Brody GS. Control of drooling by translocation of parotid duct and extirpation of mandibular gland. Dev Med Child Neurol 1977; 19: 514–7. 4. Bekkers S, Delsing CP, Kok SE, et al. Randomized controlled trial comparing botulinum vs surgery for drooling in neurodisabilities. Neurology 2019; 92: e1195–e204. 5. Bekkers S, Pruijn IMJ, Van Hulst K, et al. Submandibular duct ligation after botulinum neurotoxin A treatment of drooling in children with cerebral palsy. Dev Med Child Neurol 2020; 62: 861–7. 6. Jongerius PH, van den Hoogen FJ, van Limbeek J, Gabreels FJ, van Hulst K, Rotteveel JJ. Effect of botulinum toxin in the treatment of drooling: a controlled clinical trial. Pediatrics 2004; 114: 620–7. 7. Bekkers S, van Ulsen KJ, Adang EMM, Scheffer ART, van den Hoogen FJA. Cost-effectiveness of botulinum neurotoxin A versus surgery for drooling: a randomized clinical trial. Dev Med Child Neurol 2020; 62: 1302–8. 8. Schroeder AS, Kling T, Huss K, et al. Botulinum toxin type A and B for the reduction of hypersalivation in children with neurological disorders: a focus on effectiveness and therapy adherence. Neuropediatrics 2012; 43: 27–36. 9. Albrecht P, Jansen A, Lee JI, et al. High prevalence of neutralizing antibodies after long-term botulinum neurotoxin therapy. Neurology 2019; 92: e48–e54. 10. Benecke R. Clinical relevance of botulinum toxin immunogenicity. BioDrugs 2012; 26: e1–9. 11. Dressler D. Clinical presentation and management of antibody-induced failure of botulinum toxin therapy. Mov Disord 2004; 19(Suppl. 8): S92–S100. 12. Lange O, Bigalke H, Dengler R, Wegner F, deGroot M, Wohlfarth K. Neutralizing antibodies and secondary therapy failure after treatment with botulinum toxin type A: much ado about nothing? Clin Neuropharmacol 2009; 32: 213–8. 13. Erasmus CE, van Hulst K, Scheffer AR, et al. What could predict effectiveness of botulinum toxin to treat drooling: a search for evidence of discriminatory factors on the level of body functions or structures. Eur J Paediatr Neurol 2012; 16: 126–31. 14. Sreebny LM. Saliva in health and disease: an appraisal and update. Int Dent J 2000; 50: 140–61. 15. Bekkers S, van Hulst K, Erasmus CE, Delsing CP, Scheffer ART, van den Hoogen FJA. An evaluation of predictors for success of two-duct ligation for drooling in neurodisabilities. J Neurol 2020; 267: 1508–15. 16. Delsing CPA, Bekkers S, van Hulst K, Erasmus CE, van den Hoogen FJA. Unsuccessful submandibular duct surgery for anterior drooling: Surgical failure or parotid gland salivation? Int J Pediatr Otorhinolaryngol 2019; 123: 132–7. 17. van Hulst K, Lindeboom R, van der Burg J, Jongerius P. Accurate assessment of drooling severity with the 5-minute drooling quotient in children with developmental disabilities. Dev Med Child Neurol 2012; 54: 1121–6.

RkJQdWJsaXNoZXIy MjY0ODMw