Treatment of sialorrhea in children with Cerebral Palsy: A double-blind placebo controlled trial

1. Introduction 

Sialorrhea (drooling) is an important clinical, social and emotional issue for children and adults with neurological disease, such as cerebral palsy (CP), Parkinson’s disease (PD), and Motor neuron disease [1]. It may cause skin erosions, fungal infection, and dehydration. Drooling has been reported in 10–37% of patients with CP [2], [3]. Therefore, treatment of drooling is of significance and is part of the multidisciplinary care these patients require.

Treatment of drooling has limited success. Systemic anticholinergic medications are widely used but are not selective and have a significant side effect profile [4]. In addition to intellectual problems these patients have, significant physical problems such as poor oro-motor control hinder behavioral therapy of limited use [5]. Several surgical approaches have been performed including submandibular rerouting alone, submandibular duct rerouting with bilateral tympanic neurectomy with bilateral submandibular gland excision, and bilateral submandibular gland excision with bilateral parotid duct ligation [6], [7]. However postoperative complications such as markedly thicker secretions and major changes in the ability to swallow food limit their widespread use.

The local application of Botulinum toxin (BoNT) to the salivary glands is being increasingly reported as a safe and effective alternative to treat drooling [8], [9], [10], [11]. This controlled study aimed at evaluating the efficacy and safety of local application of BoNT for the treatment of drooling in children with CP compared to placebo control.

2. Materials and methods 

Children with CP were recruited through a local CP multidisciplinary rehabilitation center. This is a charity-based center that deals with all children with CP in the city. Physiotherapist, speech therapist and a neurologist usually evaluate the children. Patients who experienced drooling were screened by the center physician using a rating scale that assesses the severity and frequency of drooling (Table 1) [12] and those with a total score of seven or more were enrolled. Patients taking oral treatment for drooling in the last 3 months or had received BoNT injection for any other indication in the last 6 months were excluded. The study was designed so that patients would receive a dose of 100 units BoNT (Dysport) or placebo (normal saline 0.9%) in the first visit and 140U BoNT or placebo in the second visit 4 months later regardless of the response to the first injection. Thirteen patients were assigned to the control group and 11 to the study group. The caregivers signed an informed consent. The research Ethics Board of King Abdullah University Hospital and the Research Ethics Board of Jordan University of Science and Technology (JUST) approved the study.

Table 1. Rating scales for frequency and severity of drooling.

	Frequency
	1=never drools
	2=occasionally drools (not every day)
	3=frequently drools (part of every day)
	4=constantly drools
	Severity
	1=dry (never drools)
	2=mild (only lips wet)
	3=moderate (wet on lips and chin)
	4=severe (drool extends to clothes)
	5=profuse (hands, tray and objects wet)

Twenty-four children ages 21 months to 7 years were enrolled; 9 girls and 15 boys. One of the investigators assessed the severity and frequency of drooling using the previously mentioned rating scale. Another investigator injected all children using the technique described by Glickman and Deaney [13]. Each patient was given a number and a registered nurse, independent from the investigators, assigned the patients to the treatment or placebo group. The drug was diluted with normal saline to a concentration of 20U per 0.1cc. Anesthesia was not used. Each side was injected in two sites with a total of 50 Units of the drug or placebo on each side with a 10mm (30G) needle using the same volume in the syringe. Ultrasound guidance was not used. Both the investigators and the patients were blinded to which drug was given. All patients had follow up evaluation of their response and side effects at 1 month after the treatment. Anticipated side effects were explained to the caregivers before the injection including difficulty in swallowing, increase in saliva or mouth dryness or flu-like illness. A second set of injections was scheduled months after the first injection. Prior to the second injection, patients underwent another baseline evaluation using the same rating scale. The caregivers of eight children declined the second injection; six were in the placebo arm. The remaining 16 children were injected with 70 Units of the drug or placebo in each side, and the same follow up was done.

For statistical analysis, treatment and placebo groups were compared at baseline in age using Mann–Whitney U test. Gender, frequency, and severity of drooling were compared using Fishers’ Exact Test. The significance of reduction in frequency, severity, and total scores was tested using Wilcoxon Signed Rank test. Data was analyzed using Statistical Package for Social Sciences (SPSS, version 13). A p-value of less than 0.05 was considered statistically significant.

3. Results 

Children in both groups were similar in age, gender distribution, frequency of drooling, and severity of drooling. Table 2 summarizes the patients’ baseline characteristics.

Table 2. Baseline characteristics of study participants.

	Variable		Treatment N (%)	Placebo N (%)	p-Value
	Age	(Mean±S.D.) year	3.5±1.7	4.5±2.0	0.225
		Median year	3.0	4.0
	Gender	Boy	7 (63.6)	8 (61.5)	0.625
		Girl	4 (36.4)	5 (38.5)
	Frequency of drooling	Median	4	4
		Constantly drools	4 (36.4)	5 (38.5)	0.625
		Frequently drools	7 (63.6)	8 (61.5)
	Severity of drooling	Median	5	5
		Profuse	2 (18.2)	0 (0)	0.199
		Severe	9 (81.8)	13 (100)
	Median total score		9	9

Both frequency and severity of drooling did not change in 12 out of 13 children who received placebo. One child who was “frequently drool” changed to “occasionally drool”. The same child changed from profuse to mild drooling. On the other hand, the frequency of drooling declined in five children who received treatment and the severity declined in six children (one child with severe drooling and five with profuse drooling). Changes in the frequency and severity of drooling after one month of the injection are shown in Fig. 1, Fig. 2, respectively.

View Large Image Download to PowerPoint

Fig. 1. Distribution of frequency of drooling scores before and after the first injections.

View Large Image Download to PowerPoint

Fig. 2. Distribution of severity of drooling scores before and after the first injections.

The median frequency score (4), median severity score (5), and median of the total drooling score (9) did not change in placebo group. However, the median frequency of drooling score declined from 4 to 3 (p=0.034), the median severity score declined from 5 to 4 (p=0.026), and the median total score changed from 9 to 7 (p=0.027) in the treatment group. Two patients in the active treatment group experienced transient increase in drooling after the injection. The caregivers reported this side effect by telephone after two weeks of treatment, but this was not noticed during the one-month evaluation. The drooling returned to baseline eight weeks after treatment. These two patients had no benefit from treatment, and the increase in drooling was not reported after the second treatment. No other side effects were reported.

Sixteen patients received the second injection (nine in the treatment group and seven in the placebo group). Five out of seven patients who had follow up had a lower total drooling score after injection (Table 3). Three patients who did not respond to the first injection responded to the second injection. One patient in the placebo group had a decrease in the drooling score (total score decreased from 9 to 5). There was no significant difference in the median age, baseline total score and the one-month post-injection total score between patients who dropped out and patients who received second injection (p=0.490, 0.982, 0.528, respectively). Four patients lost follow up because their phones had been disconnected (two in the active treatment group and two in the placebo group. Patients in the second injection did not report any side effects. Fig. 3 summarizes the response after treatment with initial dose and high dose.

Table 3. Distribution of total drooling scores before and after each treatment.

	Patient	Gender	Age	1st baseline	After 1st injection	2nd baseline	After 2nd injection	Drug given
	1	M	5 years	8	2	8	2	BoNT
	2	M	3 years	8	8	8	8	PLB
	3	M	3 years	9	4	9	4	BoNT
	4	M	3 years	9	9	9	9	PLB
	5	M	4 years	9	9	9	9	PLB
	6	M	4 years	9	9	9	5	BoNT
	7	p	3 years	9	9	9	2	BoNT
	8	M	2 years	9	9	9	5	PLB
	9	M	3 years	8	4	8	8	PLB
	10	F	3 years	8	8	8	4	BoNT
	11	F	2 years	9	6	9	No follow up	BoNT
	12	F	3 years	9	9	9	No follow up	PLB
	13	M	8 years	8	8	8	No follow up	PLB
	17	M	3 years	9	5	9	No follow up	BoNT
	18	M	3 years	7	4	4	4	BoNT
	19	M	8 years	9	8	8	8	BoNT

View Large Image Download to PowerPoint

Fig. 3. Distribution of patients before and after each treatment.

4. Discussion 

Healthy subjects secrete between 500 to 1000ml/day of saliva. The parotid and submandibular glands produce about 90% of saliva [14]. The parotids produce the majority of saliva when stimulated, whereas the submandibulars produce the majority of saliva at rest [15]. Salivation is mediated through the parasympathetic nerves; acetylcholine is the transmitter at the parasympathetic junction. The BoNT blocks acetylcholine release at the cholinergic neurosecretory junction of salivary glands.

Placebo-controlled trials to evaluate the efficacy of BoNT in the treatment of drooling in children with CP are deficient [16]. In a placebo-controlled trial, Lin and his group evaluated the effect of BoNT-A injection into one parotid gland and contralateral submandibular gland under ultrasound guidance in thirteen patients using three methods of drooling measurement, and noted significant improvement in the treatment group lasting for 12 weeks [16]. As in our trial, this study has limited number of patients. Reid et al. studied 48 patients with different neurological disorders, and compared the effect of treatment with BoNT-A injection into each parotid and submandibular gland to no treatment in the control group and found significant difference in the drooling impact scale questionnaire in the treatment group compared to the control group [17]. However, this was an open label study.

Our case controlled study showed a decrease in drooling. The reported side effect, which was an increase of drooling, was minor and transient. This can be explained by improper placement of the injection or to local leakage of BoNT into surrounding muscles resulting in weakness of mouth closure.

In our study we elected to use rating scale [12] to measure the efficacy of treatment since it is easier for parents to judge treatment success by number of times they need to clean saliva and to notice when the effect of medicine wore off. Using dental rolls to measure efficacy of treatment of drooling was difficult in such young children for the fear of choking, and therefore it was not used. Bothwell et al. injected nine children in the parotid gland, and assessed the response by using the same rating scale we used plus weighing dental bibs and found that five patients were “moderate to good responders” [8]. However, this was an open label trial.

Jongerius and his group compared submandibular gland injection with BoNT versus scopolamine patch placed behind the ear and assessed the response by using dental rolls placed directly at the orifices of salivary glands, and they found that compared with the baseline, the mean salivary flow rate was reduced by 25% during scopolamine patch and 42% following BoNT. The response for BoNT was maintained throughout the 24-week duration of the study; four patients discontinued scopolamine patch because of side effects [9].

We chose the parotid glands to inject because of easy anatomical landmarks. Ultrasound guided injection was not used although most recent pediatric studies have used ultrasound guidance [16], [17], [18], [19], [20]. However, there are limited data on direct comparison between blind injection based on anatomic landmarks and injection with ultrasound guidance. Dogu et al. injected 15 patients with PD, eight with ultrasound guidance and seven with blind injection. He found that quantitative saliva assessments were better with ultrasound injection than blind injection throughout the follow up visits [21]. Pal and his group injected nine patients with PD with BoNT using anatomical landmarks and found that there was a marked objective reduction in secretion, and two thirds of the patients had subjective improvement in drooling [22].

We attempted to evaluate whether high dose would be more efficacious by escalating the dose up. Only nine patients received BoNT in the second set of injections, and improvement was reported in five patients including three patients who did not respond to the first injection. This suggests that higher dose might be more efficacious. However, the number is too small to establish statistical significance. Lipp et al. in a prospective, double-blind placebo controlled dose finding study injected 32 patients with different neurological disorders with placebo, 18.75, 37.5, or 75MU of BoNT (Dysport). A significant decrease of drooling was observed within the high-dose group. Medium and low doses of BoNT provoked less-pronounced effects [1].

The interventional nature of this study and the cost of medication limited our ability to recruit larger number of patients. This would have allowed us to study the dose–response effect. Lack of effect, the long distance between the rehabilitation center and our institution, and the relative discomfort associated with the procedure are some of the reasons that discouraged caregivers to come back for the second injection. The use of ultrasound guidance in our study could have improved the outcome. Comparing intraparotid injection alone versus combined submandibular gland and parotid glands injection would allow us to know the best approach for such cases.