While it is common for infants and toddlers to drool, this tendency typically diminishes through maturation of oral sensorimotor functions and is considered pathological after the age of 4 years.¹ Nevertheless, persistent anterior (i.e. visible) drooling remains prevalent in children with cerebral palsy (CP) or other developmental disabilities,² resulting in a variety of practical, physical, and social–emotional consequences that impact daily life and care.^{1, 3, 4} Moreover, some children with developmental disabilities experience posterior drooling,⁵ where saliva uncontrollably flows into the pharynx rather than spilling from the mouth, posing an increased risk of respiratory infections and significantly impacting their quality of life.¹ This emphasizes the imperative for timely and effective treatment to address drooling in these children.

Intraglandular injections of botulinum neurotoxin A (BoNT-A) are regarded as an effective treatment choice for drooling,⁶ alongside conservative treatment options, anticholinergic medications, or surgical interventions.⁷ BoNT-A works by blocking the release of acetylcholine at the neuromuscular junction within the salivary glands, thereby decreasing saliva production.⁶ Nevertheless, no consensus has been reached on the most suitable major salivary glands to target (i.e. the submandibular glands, the parotid glands, or both). Whereas the submandibular glands produce the majority of resting saliva, which has a seromucous consistency, the parotid glands substantially increase their production of serous saliva when stimulated (e.g. during the chewing of food).⁸ To address this, our saliva control team at the Radboudumc Amalia Children's Hospital adopted a stepped care approach for children treated with BoNT-A. The initial injections target the submandibular glands to diminish resting saliva production. If the response is deemed insufficient—after a shared decision-making process involving healthcare professionals, child, and caregivers—concurrent injections into the submandibular and parotid glands may be used in a subsequent BoNT-A treatment session, diminishing both resting and stimulated saliva production.⁹

As with any treatment, side effects may occur after injections with BoNT-A, specifically contributing to negative effects on oral motor function. The presumed underlying causes are associated with alterations in saliva viscosity, diffusion of BoNT-A beyond the salivary gland (i.e. leading to muscle weakness), and BoNT-A being overly effective (i.e. leading to xerostomia).^{1, 10, 11} Nevertheless, research on negative effects after BoNT-A injections is relatively limited, with a predominant focus on treatment effectiveness. Studies that specifically report on safety tend to focus primarily on adverse events rather than a broader assessment of the included children.¹² A prior study of our saliva control team assessed negative effects on oral motor function in a cohort of 209 children treated with submandibular injections.¹³ Negative effects occurred in 33%, with the majority being moderate and resolving spontaneously within 4 weeks. A study of 26 children treated with BoNT-A in both the submandibular and parotid glands reported overall worsening of eating and drinking in 2 out of 20 (10%) children,¹⁴ although this rate should be interpreted cautiously given the small sample size.

We hypothesize that, given the functional role of the parotid gland, negative effects related to eating and swallowing might be more common for concurrent submandibular and parotid injections and that problems related to an excessively dry mouth may occur more frequently considering the injection of four major salivary glands. Within this vulnerable patient population, a carefully tailored treatment approach is essential. Should data suggest an elevated rate of these negative effects, adjustments in parental education before treatment are necessary and a reassessment of our treatment strategy becomes imperative. Therefore, this study aims to evaluate negative effects on oral motor function after concurrent BoNT-A injections into the submandibular and parotid salivary glands.

METHOD

A retrospective cohort study was conducted, using data from children (aged 4–18 years) with a developmental disability, primarily neurodevelopmental, who were treated with concurrent submandibular and parotid BoNT-A injections to diminish drooling at the Radboudumc Amalia Children's Hospital between April 2004 and February 2023. Children frequently underwent multiple rounds of BoNT-A injections as part of the treatment approach for drooling during their childhood; however, this study includes only the first concurrent submandibular and parotid injection per child, with the numbers representing unique individuals. Exclusion criteria were: (1) simultaneous use of systemic anticholinergic medication (i.e. between the date of injection and 32-week follow-up); (2) previous surgical saliva control treatment (e.g. submandibular duct relocation, submandibular gland excision, salivary duct ligation); and (3) no standard 8-week follow-up information available.

All included children received the injections as part of routine clinical care, involving an ultrasound-guided injection of BoNT-A (Botox, Allergan, Nieuwegein, the Netherlands) administered in the submandibular glands (bilaterally) and parotid glands (either unilaterally or bilaterally), while under general anaesthesia.¹⁵ Generally, a dose of 25 U BoNT-A was used. Since our team applies a stepped care approach—starting with bilateral submandibular injections and opting for these concurrent submandibular and parotid injections only if the initial treatment is insufficient (with a few exceptions)—most children included in the study were not naive to BoNT-A treatment.

Standard follow-up appointments with a speech-language therapist from our team were scheduled at 8 weeks and 32 weeks post-injection, with the 8-week follow-up increasingly conducted via telephone in recent years. Three speech-language therapists are part of our saliva control team, one of whom has been involved since the start of data collection. Both speech-language therapists that joined the team later were familiarized with the assessment procedures to ensure consistency in measurement.

The local research ethics committee of the Radboudumc deemed the study exempt from ethical review (reference: 2023–16435). As this study comprised an analysis of data collected for routine clinical care, the committee waived the need to obtain informed consent for children no longer receiving active care (i.e. last visit to the outpatient clinic more than 7 years ago and/or deceased). For the remaining children, informed consent was obtained from parents and/or caregivers.

RESULTS

DISCUSSION

To our knowledge, this study describes the first comprehensive examination of negative effects on oral motor function among children treated with concurrent submandibular and parotid gland injections within a cohort of this size. Negative effects were reported for one-third of the included children, primarily manifesting as eating-related problems and difficulties in saliva swallowing. Encouragingly, the majority of these problems were classified as mild and resolved within 4-weeks post-injection, with only two cases lasting longer than 8 weeks.

The overall reported rate for any negative effects after concurrent submandibular and parotid BoNT-A injections differs considerably across published literature, generally ranging from 0% to 40% and occasionally reaching up to 78%.²¹ Nevertheless, it is important to note that differences in treatment protocol and outcome definitions (e.g. adverse events, side effects, complications, secondary effects) may have influenced these rates, along with differences in characteristics of the study population.²¹ For example, Berweck et al. recently reported adverse events in 27 cases (18.2%), but these mostly consisted of non-treatment-related respiratory infections.²² Our focus was on expected negative effects of treatment, and it is not always evident whether these are encompassed in the analyses of other studies.

Generally, the negative effects rate in the current study mirrored our reference cohort treated with submandibular injections. However, these cohorts differed in several distinctive clinical characteristics, particularly in the proportion of children with CP, a previously identified risk factor for negative effects.¹³ Moreover, a majority of included children had prior exposure to BoNT-A in the submandibular glands because of our stepped care approach, and a step-up to concurrent submandibular and parotid injections will likely not have been indicated in those with severe negative effects to initial injections. It remains uncertain whether this may have impacted susceptibility to negative effects in the current cohort. Additionally, it is worth considering that parents with prior experience with this treatment may have taken additional precautions during this treatment session that we were unaware of, got used to certain side effects, or were less worried, further influencing outcomes. These considerations indicate that the negative effect profile of concurrent injections in children previously treated with submandibular injections should not simply be assumed to mirror that of stand-alone concurrent injections. Establishing the generalizability of our findings in a treatment-naive cohort is therefore crucial.

Although negative effects primarily manifested as problems with eating, the prevalence of eating-related problems did not appear to differ from our reference cohort, in contrast to our hypothesis. However, differences in feeding methods between the two cohorts should be taken into account, with 70.4% primarily fed orally in the current cohort compared to 83.7% in the reference cohort, meaning that the population of children at risk for negative effects during mealtimes was smaller. Reid et al. included a population of children with a higher prevalence of oral feeding (84.6%), similar to our reference cohort, and found worsening of eating and drinking in only 10% of children.¹⁴ However, this lower proportion may be attributed to factors including the relatively small sample size or the method of assessing secondary effects, based on a questionnaire at 4-weeks post-treatment. A certain correlation between treatment effect and the occurrence of negative effects was hypothesized before the study, associated either with technical failure of the injections (i.e. reflected by treatment non-response) or with excessive effectiveness of the treatment and/or the thickening of saliva¹⁰ (i.e. reflected by treatment response). Although no difference in the overall negative effects rate could be identified, children classified as responders more often appeared to have moderately severe negative effects rather than mild negative effects compared to non-responders, potentially reflecting this second hypothesis.

Interestingly, no severe cases of negative effects were found in the current cohort, a notable contrast to the reference cohort as well as findings in existing literature. For example, Khan et al. reported adverse events in 15 children (33.3%), with five experiencing severe complications.²³ Similarly, a study by Chan et al. revealed adverse events in 19 children (15.8%), with major complications in five cases.¹² This discrepancy may be a testament to the value of our stepped care approach, as children experiencing severe negative effects after submandibular BoNT-A injections are unlikely to receive concurrent submandibular and parotid injections. Analogous to drug titration, this approach may enable the identification of those susceptible to severe side effects at a lower BoNT-A dose (i.e. submandibular injections), potentially preventing more severe negative effects that might arise when submandibular and parotid injections would be given. However, it should be acknowledged that this discrepancy could also be a consequence of prior exposure to BoNT-A in general, rather than a distinct difference between submandibular and concurrent submandibular and parotid injections.

To better inform decision-making, it is crucial to understand both the expected treatment effects and potential side effects of concurrent BoNT-A injections for drooling. Nevertheless, it is important to recognize that response rates can vary depending on the criteria used and may not fully reflect individual satisfaction or perceived effectiveness.²⁴ Likewise, negative effects may not always be perceived as outweighing the benefits of treatment but should rather be considered as important aspects to be anticipated. Therefore, our study aimed to provide caregivers with realistic expectations about potential negative effects and their severity, ensuring they are well informed before treatment and allowing them to make necessary adaptations to support the child if changes in oral motor function occur. Moreover, it remains important to apply an individualized treatment approach, in which the benefits and potential negative effects of treatment are continuously weighed for each individual child.

The strength of our study lies in the large sample size and comprehensive collection of information regarding negative effects, facilitated by routine follow-up evaluations where documenting any negative effects is standard practice. However, several limitations should be acknowledged. First, mild negative effects may not have been consistently disclosed by parents during follow-up evaluations or may not have been deemed substantial enough to be documented in the medical records. Second, it has been hypothesized that negative effects may occur more frequently in children exposed to multilevel BoNT-A injections, where intraglandular injections are combined with injections into the skeletal muscles.¹³ We were unable to explore this in the current cohort. Third, the data for this study were collected over a 20-year period, involving three speech-language therapists in follow-up. However, a thorough onboarding process was implemented to standardize follow-up procedures, thereby maintaining consistency and minimizing variations in reporting negative effects after BoNT-A injections.

In conclusion, this study sheds light on the frequency, severity, and duration of negative effects on oral motor function after concurrent submandibular and parotid gland BoNT-A injections, in children predominantly treated with submandibular injections initially, providing valuable insights to inform parents and children before treatment. Our findings suggest that clinicians can confidently consider adding parotid injections to the treatment protocol when submandibular injections prove ineffective, given the comparable negative effect profile in this specific setting. While careful interpretation of these results is warranted, considering that most children were previously exposed to intraglandular BoNT-A, our findings also prompt the hypothesis that a stepped care approach may be beneficial. Specifically, it may prevent severe negative effects after concurrent submandibular and parotid injections by selectively excluding children at risk for these consequences at an earlier stage. Further investigation is needed to elucidate the nuances in this observed pattern.