Drooling, saliva production, and swallowing in cerebral palsy
Senner, Jill E
Drooling, ‘spilling of saliva from the mouth onto the lips, chin, neck, and clothing’ (Brodsky 1993, p389), normally occurs in infants and young children, particularly when a child is learning a new motor skill or cutting a new tooth. By the time typically developing children are 24 months old, they should have the ability to perform most activities without drooling (Morris and Klein 1987). It has been estimated that drooling abnormally persists in 10 to 38% of individuals with cerebral palsy (CP; Johnson and Scott 1993). Consequences of drooling include irritated facial skin, unpleasant odor, increased oral and perioral infections, hygiene problems, and dehydration (Cotton and Richardson 1981, Harris and Purdy 1987, Lew et al. 1991). Severe drooling often necessitates frequent clothing changes and can cause damage to books, electronic equipment, and other educational materials. The most unfortunate consequence of drooling may be social isolation. Because drooling is unsightly and produces an unpleasant odor, people may avoid individuals who drool and physical contact may be reduced. Isolation, in turn, can have devastating effects on self-esteem (Blasco and Allaire 1992).
Why some individuals with CP drool and others do not is not entirely clear. Some researchers have suggested that drooling is caused by disordered oral motor functioning (Burgmayer and Jung 1983, O’Dwyer et al. 1989, Lespargot et al. 1993). Others have noted that ‘Drooling might be due to hypersalivation and/or an insufficient mechanism for the removal of saliva’ (Ekedahl et al. 1974, p 78).
Researchers thus far have determined that children who drool have increased difficulty forming a bolus (Ekedahl et al. 1974), reduced lip closure, slightly less intraoral suction, and more oral residue after the swallow (Lespargot et al. 1993). Significant negative correlations have been found between drooling and sucking ability, drooling and chewing ability, and drooling and swallowing. Significant positive correlations have been found between drooling and poor head control, reduced ability to voluntarily control the lips, reduced ability to voluntarily control the tongue, and reduced ability to voluntarily control the jaw (Van De Heyning et al. 1980). Other investigators have found reduced intraoral sensitivity (Weiss-Lambrou et al. 1988), reduced frequency of spontaneous swallowing (Sochaniwskyj et al. 1986), esophageal stage abnormalities (Ekedahl et al. 1974), dental malocclusion (Franklin et al. 1996), and poor coordination of obicularis oris and masseter activity (Sochaniwskyj et al. 1986) in children who drool.
Little is known about saliva production in children with CP. Davis (1979) found significantly lower stimulated parotid flow rates in children with moderate to severe head and neck involvement due to spastic CP than in control participants. Chiat and Kessler (1979) noted that the excised submandibular glands of drooling patients with CP appeared hypertrophied and enlarged.
Recently, Tahmassebi and Curzon (2003) attempted to address the role saliva production plays in drooling in children with CP They concluded that children with CP who drooled did not produce excess saliva; however, their data must be interpreted with caution because of significant methodological limitations. The authors compared children with CP with typically developing age- and sex-matched peers, but ‘the method and duration of saliva collection in the control group was chosen to be different from those of CP individuals’ (p 109). In the control participants, the researchers used a ‘draining method’, an unstimulated method of whole saliva collection. With the study participants, on the other hand, they used a chin cup described by Sochaniwskyj in 1982. Each participant participated in only one 15-minute collection system. It should be noted that the system described in Sochaniwskyj’s study is described as a method of quantifying drooling, not whole saliva production. The chin cup only measures saliva spilled and does not account for saliva remaining in the oral cavity. Furthermore, Sochaniwskyj noted, ‘Based on experience from this study at least five 30-minute sessions must be conducted with a participant in order to obtain a true estimate of drooling rates and variability’ (Sochaniwskyj 1982, p 607).
Unfortunately, despite the fact that limited data exist about the amount of saliva produced by individuals with CP who drool, many drooling treatments, including surgical management, medications, and injections of botulinum toxin A, are designed to reduce the amount of saliva produced. The purpose of this study was to help us define those factors that influence drooling.
Participants and controls were recruited from several different schools, therapy centers, and residential facilities serving individuals with CP in the Chicago area. Letters were sent to professionals who in turn referred appropriate candidates after securing informed consent. Three groups were examined in this investigation: (1) 14 participants with CP who drooled (CP+) ; (2) 14 participants with CP who did not drool (CP-); and (3) 14 typically developing participants (CTRL). The study protocol was approved by the Institutional Review Board of Northwestern University. Participants (six females, eight males) ranging in age from 7 years 11 months to 18 years 2 months (mean 11y 8mo, SD 3y 5mo) with a confirmed medical diagnosis of spastic CP served as study participants (see Table I). All participants with CP+ met the following criteria: (1) exhibited severe to profuse and frequent to constant drooling (as determined by questionnaire); (2) no medication to reduce saliva production for at least 72h; (3) no history of head or neck radiation; (4) no ulcerations in the oral cavity; and (5) no surgical procedure to reduce drooling. Each parent was required to complete a questionnaire that contained questions about drooling seventy and frequency (Thomas-Stonell and Greenberg 1988), presence of factors related to drooling like open mouth posture and poor head control (Van De Heyning et al. 1980), presence of factors related to saliva production (e.g. medications, gastroesophageal reflux), cognitive ability, feeding and swallowing ability, and severity of motor involvement (Gisel and Alphonce 1995).
On the questionnaire, each drooling severity level and frequency related to a number on an ordinal scale from one to five (e.g. moderate has a score of three, severe has a score of four, and profuse has a score of five). Severity and frequency numbers were added for each participant to yield a drooling quotient. Participants with a drooling quotient score of 6 to 10 were included as participants. Fourteen age- and sex-matched controls were included in each of two groups: (1) participants with CP+ who were dry to mild and never to infrequent droolers, as determined by questionnaire (CP-); and (2) typically developing peers (CTRL). The cumulative drooling quotient scores for control participants were in the range two to four.
The mean difference in age between CP+ and CP- groups was 7.85 months, and the mean difference in age between CP+ and CTRL groups was 6.79 months. Control participants also met criteria 2 through 5 above.
Before the evaluation session, each participant’s relevant medical records were obtained to confirm the diagnosis of CP+ and provide any information on past evaluation or treatment of drooling or swallowing. In addition to completing the above-mentioned questionnaire, the Pediatric Evaluation of Disability Inventory (PEDI) was also administered. The PEDI is a clinical assessment instrument that assesses children’s functional abilities in a variety of areas including self-care, mobility, and social function (Feldman et al. 1990; Haley et al. 1991, 1992). Information from the questionnaire and PEDI, with a videotape of the evaluation session, were used to assess severity of CP according to the Gross Motor Function Classification System (GMFCS) for CP (Palisano et al. 1997).
A second researcher independently assigned GMFCS scores to nine of the participants to establish interrater reliability. Reliability analysis of these data revealed good interobserver agreement (Spearman’s rho correlation=0.94, p
Questionnaire information and observation of participant were also used to determine severity of dysarthria as measured on a five-point scale (Yorkston et al. 1999). Scores were assigned as follows: (1) no detectable speech disorder; (2) obvious speech disorder with intelligible speech; (3) reduction in intelligibility; (4) natural speech supplemented with augmentative and alternative communication techniques; (5) no functional speech.
The evaluation session itself consisted of three sections. On the day of the evaluation session, parents of participants were instructed not to allow their child to eat or drink anything or chew gum for 2 hours before the procedure. In addition, if a participant developed a cold, flu, or upper respiratory infection, parents were instructed to reschedule.
To quantify the amount of saliva the participants produced, the Saxon test (Kohler and Winter 1985), a simple gauze-chewing procedure, was performed by each participant. Participants with large oral cavities used a 10.16cm × 10.16cm sterile piece of gauze (4inch × 4inch Curity, 12-ply, Kendall Co, Boston, MA) that was folded twice at 90° angles and placed in a plastic container. A 7.62cm × 7.62cm (3inch × 3inch) piece of gauze was substituted in younger participants or in participants with smaller oral cavities. The gauze and plastic container were weighed using a Mettler P-1200 balance (Mettler, Columbus, OH). Before each evaluation session, the balance was calibrated with a 1mg weight. Participants were asked to swallow to remove any preexisting oral fluid. All participants and control participants then had their oral cavities suctioned with a portable oral suction unit fitted with sterile suction tubing and a straight Yangkauer sterile suction handle. Saliva was collected by having each participant chew on the gauze for exactly one minute ‘as if chewing on food’. To avoid the risk of potential aspiration of the gauze, a string of dental floss was threaded through the gauze and held by the examiner while the participant was chewing. After the chewing procedure, the gauze was replaced in the same plastic container and weighed. The amount of saliva produced was determined by subtracting the original weight from the weight obtained after chewing. During the Saxon test, each participant wore a pre-weighed disposable Sani-Tab Chain-Free Polyback dental towel (Crosstex, New York). If any saliva was lost from the mouth, this bib was used to catch and/or wipe it. After the gauze-chewing procedure was completed, the bib was weighed and the quantity of saliva in the bib was added to the quantity in the gauze to yield the total amount of saliva produced.
To determine whether multiple trials of the Saxon test might be needed, the variability in data from the Kohler and Winter (1985) study were subjected to Spearman ranked correlations. Ten of their participants were tested on three occasions, 7 to 14 days apart. The mean correlation was calculated (Spearman’s rho correlation=0.90). Mulligan et al. (1995) also found test-retest reliability of this measurement to be high (r=0.80, p
Frequency of spontaneous swallowing during a 20-minute period was determined using concurrent cervical auscultation and videotaped observation of laryngeal elevation. Cervical auscultation is a non-invasive technique for listening to the sounds of pharyngeal swallow. This technique has been used to assess adults and children for clinical and research purposes (Vice et al. 1990, Takahashi et al. 1994, Zenner et al. 1995). Each participant’s neck was cleaned with alcohol, and a tie-clip microphone was affixed to the midline of the neck around the level of the cricoid cartilage using a double-sided electrode washer (Cichero and Murdoch 2002). The microphone was connected to the audio input channel of the video camera. The camcorder was positioned at a level slightly lower than each participant’s neck and angled upward to allow maximum visualization of laryngeal elevation. During this 20-minute period, the participants were each shown the same children’s video to help distract them from concerns with the equipment as well as to provide an opportunity to observe swallowing during a familiar childhood activity (i.e. watching television). The videotapes of the evaluation session were reviewed by the principal investigator and 25% were reviewed again to establish intraobserver reliability. Reliability analysis of these data revealed good intraobserver agreement (r=0.9959, p
While frequency of swallowing was being monitored, the amount of drooling was measured by using the bib-weighing technique. A disposable Sani-Tab Chain-Free Polyback dental towel was placed in a plastic container and weighed. Before the bib-weighing procedure, each participant’s mouth and chin were wiped with a paper towel to remove any drool present before the procedure. Next, each participant wore the bib for exactly 10 minutes. After the 10-minute period, any saliva on the lips or chin was wiped with the bib. The bib was replaced in the plastic container and weighed. The amount of drool produced was determined by subtracting the original bib weight from the weight obtained after the 10-minute period.
Finally, each participant was given the Test of Nonverbal Intelligence-3 (TONI-3). TONI-3 consists of 50 questions, and administration takes about 15 minutes. Participants were required to look at a series of pictures and select appropriate responses with pointing or eye-gaze. No verbal responses were required. Standard scores were computed for the TONI-3.
Drooling severity as measured by questionnaire (drooling quotient) and the bib weighing technique were compared using Spearman’s rho correlation. Results indicated that there was a positive correlation between caregiver reports of drooling severity and bib weight during the examination session (Spearman’s rho correlation=0.604, p
All of the participants tolerated the Saxon test well, without any overt signs of anxiety such as protest or crying; however, two participants in the CP+ group and three participants in the CP- group gagged on the gauze owing to intraoral hypersensitivity. None of the typically developing participants gagged during the Saxon test. The impact of gagging on saliva production is not clear. On this test, differences were not determined to be statistically significant by an analysis of variance (ANOVAF^sub (2,39)^=2.842, p=0.071; see Fig. 1). No correlation was found between amount of saliva produced and amount drooled (r=0.245). Lip incompetence during chewing, as evidenced by gauze protruding out of the mouth, was observed in all of the CP+ participants and 11 of the 14 participants in the CP- group during the Saxon test. All of the typically developing control participants were able to achieve complete lip closure around the gauze. The control participants also demonstrated the ability to form a bolus with the gauze and, at the end of the 60 seconds, the participants spat out rounded gauze balls. All of the participants in the CP+ group and 11 of the participants in the CP- group were unable to form a bolus (e.g. the gauze was still square in shape), suggesting poor oral coordination.
Both groups of participants with CP tended to swallow less frequently than typically developing participants, with participants in the CP- group swallowing more frequently than participants in the CP+ group (Fig. 2). However, an ANOVA revealed no statistically significant difference between the three groups (F^sub (2,39)^ = 1.8274, p=0.1743).
An ANOVA conducted on the PEDI functional skills average scores indicated significant differences between the three groups (F^sub (2,39)^=23.522, p
Likewise, an ANOVA conducted on the TONI-3 scores revealed statistically significant differences between the three groups (F^sub (2,39)^=31.761, p
In the CP+ group, one participant had a GMFCS score of 1, four participants received a score of 2, two received a score of 3, three received a score of 4, and four participants received a score of 5. In the CP- group, four participants received a GMFCS score of 1, four received a score of 2, two received a score of 3, three received a score of 4, and one participant received a score of 5. A Spearman’s rho correlation indicated that gross motor severity was not correlated with drooling severity as measured by the questionnaire (drooling quotient).
Eight of 14 of the participants in the CP+ group used some formal augmentative and alternative communication system (e.g. picture boards, sign language, voice output communication aids) whereas none of the participants in the CP-group was reported to use any forms of augmentative and alternative communication. A Fisher’s exact test indicated significant differences between the groups (p-exact
On the questionnaire, caregivers were also asked whether or not participants had experienced seizures. Eight out of 14 participants in the CP+ group had been diagnosed with seizure activity. Only 1 of 14 participants in the CP- group had been diagnosed with seizure activity. A Fisher exact test indicated significant differences between the groups (p-exact=0.006086) for this item. All of the participants who experienced seizures were managed with anticonvulsant medications including carbamazepine, phenobarbital, divalporex sodium, and clonazepam.
Results of the present study suggest children with CP who drool have poorer functional skills scores, lower nonverbal intelligence scores, more severe oral motor involvement, and a tendency to swallow less frequently than children with CP who don’t drool. Participants with CP did not produce excess saliva, and no correlation was found between amount of saliva produced and amount drooled, suggesting that hypersalivation is not one of the factors responsible for drooling in this group.
As previously noted, data from this study indicate that participants with CP do not produce excess saliva. However, in this study, there was a motor component (i.e. chewing) involved in the Saxon saliva collection, that may have put participants with CP at a disadvantage. Despite methodological limitations, these results are consistent with those of Davis (1979), who found significantly lower stimulated parotid flow rates in children with moderate to severe head and neck involvement due to CP than in control participants. The results are also consistent with those of Tahmassebi and Curzon (2003), who found that children who drooled did not produce excess saliva. Furthermore, adult patients with Parkinson’s disease have also been demonstrated to produce significantly less unstimulated saliva than age- and sex-matched control participants (Bagheri et al. 1999), suggesting that drooling in these patients is related to swallowing difficulties rather than hypersalivation. It is important to note that both the present study and Davis (1979) measured stimulated saliva production. Future studies using unstimulated collection techniques would be helpful in determining factors contributing to reduced saliva production.
Furthermore, in the present study no correlation was found between stimulated whole saliva production and amount drooled. These findings are of particular importance because many drooling treatments, including surgical management, medications, and injections of botulinum toxin A, are designed to reduce the amount of saliva produced. Although most studies have not termed the post-treatment reduction in salivary flow ‘xerostomia’ as such, reduction in salivary flow has not been without consequence. In the literature, several studies have reported significant increases in dental caries after surgical intervention (Arnrup and Crossner 1990, Hallett et al. 1995).
Furthermore, it has been estimated that about a third of children with clinically significant drooling demonstrate evidence of gastroesophageal reflux (Heine et al. 1996). In this study, three of 14 of participants in the CP+ group were found to have gastroesophageal reflux, as determined by review of medical records and analysis of the intake questionnaire. Saliva plays an important role in the defense against the acid-induced esophageal mucosal injury that can occur with gastroesophageal reflux. In children who drool, constant saliva loss may already impair clearance of refluxed gastric acid from the esophagus, which in turn may perpetuate esophageal dysmotilityandesophagitis (Heine et al. 1996). Further reduction in saliva production caused by drooling treatments might only serve to exacerbate this situation.
This is not to say that reducing saliva is inappropriate. Many researchers have reported positive results in their patients following surgical procedures as determined by subjective rating scales. (O’Dwyer et al. 1989, Ethunandan and Macpherson 1998, Mankarious et al. 1999). For example, surgical procedures have been effective in eliminating episodes of aspiration pneumonia secondary to aspirated salivary secretions (Klem and Mair 1999). Nonetheless, it would seem that using more stringent candidacy criteria for saliva reduction treatments would be prudent. As a minimum, before treatment, a comprehensive examination of oropharyngeal swallowing as well as gastroesophageal reflux should be completed. The need for saliva reduction procedures would best be determined by an interdisciplinary team. The team could also monitor and possibly prevent potential sequelae with appropriate medical management. It has been suggested that team management using an otolaryngologist, a speech-language pathologist, and a pediatric dentist works well for children with developmental disabilities who drool (Crysdale 1992). Because of the medically complex nature of these children, a pediatric physician could also play an important role.
Although the data did not achieve statistical significance, there was also a tendency for participants with CP to swallow less frequently than typically developing participants, with participants in the CP- group swallowing more frequently than participants in the CP+ group (Fig. 2). These results are consistent with a previous study conducted by Sochaniwskyj et al. (1986, p 871) which found ‘The nondrooling group swallowed at approximately 75% the rate of normals, while the drooling group swallowed at 45% the normal rate’. Although Sochaniwskyj et al. (1986) concluded that drooling in children with CP was caused by ‘both inefficient and infrequent swallowing’, correlation does not imply causation. Previous studies have established that both saliva flow rate and volume are important variables in determining spontaneous swallowing frequency (Kapila et al. 1984, Rudney et al. 1995, Pouderoux et al. 1996). It is still unclear whether reduced swallowing rates cause drooling or whether reduced swallowing rates are the result of reduced saliva production coupled with the anterior loss of saliva that occurs in drooling. Researchers have suggested that ‘posterior drooling’, or pooling of saliva in the pharynx, might occur in patients with CP who drool (Blasco and Allaire 1992); however, so far, there has only been indirect evidence of this (e.g. wet, gurgling vocal quality). Further studies of children with CP who drool are needed, to investigate the amount of saliva reaching the pharynx and what happens to that saliva once it reaches the pharynx.
It is clear that drooling is caused by multiple factors including poor oral motor abilities, poor intraoral sensations, and dysphagia. Results of the present study indicate that hypersalivation is not one of the factors responsible for drooling in children with CP Because of the complex nature of drooling in children with CP, each child’s individual strengths and areas of need should be assessed by a multidisciplinary team and treatment should be determined on a case-by-case basis.
Although this study suggests that children who drool have more severe forms of CP and a higher probability of having a lower IQ, this does not mean that they cannot benefit from treatment. Treatments aimed at increasing swallowing frequency and efficiency as well as treatments to improve impaired head and neck control may be fruitful endeavors. Characteristics of the population of children with CP who drool, particularly poorer oral motor control and generally lower cognitive skills, make treatments involving following complex directives or completing oral exercises difficult. However, not all treatments require active patient participation. Recent reports in the literature about treatments such as the use of intraoral appliances (Inga et al. 2001) and tongue acupuncture (Wong et al. 2001) may hold promise and should be investigated further.
Accepted for publication 23rd March 2004.
This work was partly supported by the Dissertation Year Grant from the University Research Grants Committee at Northwestern University.
Copyright Mac Keith Press Dec 2004
Provided by ProQuest Information and Learning Company. All rights Reserved.