SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

Aim  To establish a prevalence estimate for drooling and explore factors associated with drooling in a population sample of children with cerebral palsy (CP) aged 7 to 14 years living in Victoria, Australia.

Method  A self-report questionnaire was used to collect data on drooling from parents of children born between 1996 and 2001, and registered with the Victorian Cerebral Palsy Register.

Results  A total of 385 children (231 males, 154 females; mean age 10y 9mo [SD 1y 7mo], range 8–14y) were studied. The clinical type and distribution of CP were spastic (341), ataxic (16), dyskinetic (17), hypotonic (10), and unknown (1). Distribution in Gross Motor Function Classification System (GMFCS) levels was I (103), II (98), III (52), IV (63), V (61), and unknown (8). After adjustment for topographical pattern of motor impairment and GMFCS level, 40% were reported to have experienced drooling between 4 years of age and the time of completing the questionnaire. A significantly higher prevalence of drooling was found in children with poor gross motor function and in those with more severe presentations of CP, including poor head control, difficulty with eating, and inability to sustain lip closure (p<0.001 for each). Drooling was shown to be significantly associated with both intellectual disability and epilepsy in this group of children (p<0.001 for both).

Interpretation  With a prevalence of 40%, drooling is an important comorbidity in CP. It was considered severe in 15% of children. Poor oromotor function was associated with drooling and could be the target of interventions for this under-researched problem.


Abbreviation
VCPR

Victorian Cerebral Palsy Register

What this paper adds

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References
  •  After adjustment for level of gross motor function and topographical pattern, 40% of 7- to 14-year-old children with cerebral palsy were reported to have a drooling problem.
  •  A non-spastic motor type, quadriplegic topographical pattern, lack of head control, poor gross motor function, epilepsy, intellectual disability, lack of speech, anterior open bite, and feeding difficulties predicted drooling in this group of children.

Cerebral palsy (CP) refers to a group of disorders of the development of movement and posture, causing activity limitation, that are attributed to non-progressive disturbances that occurred in the developing fetal or infant brain.1 It is a common neurodevelopmental disability, with a reported birth cohort prevalence of 1.5 to 2.5 per 1000 live births.2 The type of motor disorder in CP is commonly classified as spastic, dyskinetic, ataxic, or hypotonic and, although spasticity is the predominant motor type in approximately 90% of individuals, mixed motor types are frequently identified.3 Spasticity can be further described as having a quadriplegic, hemiplegic, or diplegic topographical pattern.4 Gross motor function varies from mild impairment, where children are able to walk independently, to more severe impairment, where children use a wheelchair for mobility and have restricted ability to maintain antigravity head and trunk postures.5,6 Approximately 50% of children with CP have some degree of cognitive impairment, and seizures and impairment of sensory modalities such as hearing, vision, pain, and touch are common.7 Oromotor dysfunction, communication impairment, and excessive drooling are also often reported in children with CP.8,9

Although early drooling may be consistent with normal oromotor development, drooling is considered abnormal when it persists beyond 4 years of age.10 For individuals with CP, drooling can be a significant social problem, potentially leading to loss of self-esteem and social isolation,11 and excessive drooling can substantially increase the work demand placed on carers.12 Abnormal oromotor tone and poor voluntary control of the head, lips, tongue, and jaw may all interfere with the ability to swallow saliva.13–15 This may be compounded by other factors that contribute to persistent drooling in CP such as delayed dental maturation and dental malocclusion.16,17

In the few studies that have reported on the prevalence of drooling in association with CP, figures have varied between 22% and 78%.18–23 These prevalence estimates vary widely and are mostly not generalizable to all children with a diagnosis of CP because they have largely been based on clinical19,22,23 or specialist school18,21 samples. Previous studies have also used variable age ranges, definitions of drooling, and measurement methods. One recent study using a yes/no question in a population cohort of 1357 children with CP in Northern Ireland estimated the prevalence of drooling to be 22%.20 This estimate is yet to be verified in other CP cohorts. The aim of this study was not only to report a prevalence estimate for drooling in a population cohort of children with CP aged 7 to 14 years, but also to describe associations between drooling and other clinical characteristics within this group.

Method

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

The study was conducted at La Trobe University and the Royal Children’s Hospital, Melbourne, Australia. Ethics approval for the study was obtained from both institutions. Completion of the questionnaire by a family member deemed to be implied consent to participate in the study.

Study participants

Eligible children were identified from the Victorian Cerebral Palsy Register (VCPR). The VCPR was established in 1986 and has subsequently registered a mean of 100 new cases of CP per year since 1970 from a denominator population of approximately 65 000 live births in Victoria.7 Multiple sources are used to identify new cases, and information on each identified case is collected from medical records, clinicians, and families. Collected data include demographic and birth details, and clinical characteristics such as motor type, topographical pattern of spasticity, functional ability, the presence of birth defects, a diagnosis of epilepsy, and intellectual, hearing, visual, and communication impairments.

To be included in this study, children were required to be living in Victoria, born between 1996 and 2001 inclusive, and registered before 1 July 2010. These birth years were chosen because they spanned a period of relative dental stability and because poor saliva control was thought unlikely to improve after the age of 7 years by the normal process of maturation. Of 722 families of children who met the inclusion criteria, 522 (72%) were able to be contacted within the ethical guidelines of the VCPR and were sent the study questionnaire. Non-responders to the first mailout were sent a follow-up reminder letter after 2 weeks. Most families (n=220) who had not responded to the second letter within 2 weeks were contacted by telephone by the second author (JM) or VCPR staff. Many families chose to complete the questionnaire by telephone.

Study questionnaire

A 16-question self-report questionnaire was constructed for the study. Face and content validity was supported by review of the items by professionals and families at our Saliva Control Clinic, but no formal psychometric testing was undertaken. The questionnaire was designed to elicit information on demographic characteristics, the presence of comorbidities, dentition, and level of gross motor function using the family-report version of the Gross Motor Function Classification System (GMFCS).5 The GMFCS has been demonstrated to be valid,24 stable,25 and highly reliable when used by parents.26 To assess oromotor function, carers were asked to categorize their child’s communication, eating, and lip seal abilities. They were also asked to indicate whether drooling had been a problem for their child beyond 4 years of age. The definition of drooling used in the questionnaire was ‘an unintentional loss of saliva from the mouth’.15 Where drooling had been identified as a problem, information was requested about past and current medical interventions. The parents’ perception of the severity of drooling was assessed using two items from the Drooling Impact Scale.27 This scale consists of a set of ten items, each measured on a scale of 1 to 10. Individual totals were calculated for items rating the severity and frequency of drooling, and classified as mild, moderate, or severe drooling based on cut-offs at 6 and 14 of a possible 20.

Statistical analysis

To identify potential participation bias, VCPR data were used to compare demographic details and clinical characteristics between participating and non-participating children using χ2 analyses. The prevalence of drooling was calculated for the children whose parents had completed the questionnaire. In addition, an adjusted population prevalence was calculated by weighting on GMFCS level and topographical pattern, thus minimizing sample bias. To assess the strength of association between drooling and the chosen demographic and clinical variables, χ2 analyses were performed to compare children with and without a problem with drooling. All data collected were analysed using Stata 11 (StataCorp 2009, College Station, TX, USA). The significance level was set at 0.05 for all tests.

Results

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

Of 522 families who were invited, 385 (81%) agreed to participate in the study. The questionnaire was completed by 362 families, either in writing or by telephone, and a further 23 provided a response only to the question about drooling. This latter group was included to determine an estimate of drooling prevalence but was excluded from subsequent analysis. Characteristics of the participating children (231 males, 154 females, mean age 10y 9mo [SD 1y 7mo]) are shown in Table I where they are compared with non-participating children from the VCPR. There was a statistically significant and clinically meaningful difference between participating and non-participating children in the degree of impairment, with participating children more likely to have poorer gross motor function, a quadriplegic pattern of motor impairment, epilepsy, intellectual impairment, lack of speech, and a gastrostomy tube in place.

Table I.   Comparison of participating and non-participating children
CharacteristicNon-participants (n=337)Participants (n=385)
  1. GMFCS, Gross Motor Function Classification System.

Mean age at 1 January 2010, y (SD)11.0 (1.7)10.9 (1.7)
Sex, n (%)
 Male194 (57.6)231 (60.0)
 Female143 (42.4)154 (40.0)
Motor type, n (%)
 Spastic289 (88.1)341 (88.8)
 Ataxic25 (7.6)16 (4.2)
 Dyskinetic7 (2.1)17 (4.4)
 Hypotonic7 (2.1)10 (2.6)
 Unknown91
Topographical pattern, n (%)
 Hemiplegia125 (38.1)121 (31.5)
 Diplegia101 (30.8)110 (28.7)
 Quadriplegia102 (31.1)153 (39.8)
 Unknown91
GMFCS, n (%)
 Level I129 (39.9)103 (27.3)
 Level II102 (31.6)98 (26.0)
 Level III26 (8.1)52 (13.8)
 Level IV39 (12.1)63 (16.7)
 Level V27 (8.4)61 (16.2)
 Unknown138
Comorbid factors, n (%)
 Epilepsy86 (26.6)127 (33.3)
 Intellectual disability112 (37.7)177 (48.1)
 No useful speech64 (21.1)116 (31.1)
 Gastrostomy feeding required21 (6.2)37 (9.6)

Prevalence of drooling

During the period from 4 years of age to the time of completing the questionnaire (aged 7–14y), 156 of 385 (40.5%) children were reported to have a drooling problem. After adjustment for GMFCS levels and topographical patterns, the prevalence dropped to 39.6% (95% confidence interval 33.8–45.4). The estimated prevalence for each combination of GMFCS level and topographical pattern is shown in Table II. Of the respondents who reported that their child had experienced a problem with drooling beyond 4 years of age, 147 completed the Drooling Impact Scale. Drooling was deemed very mild for 39 (10%) children (a rating of 6 or less out of 20 for severity and frequency of drooling). Of the remaining 108 children with significant drooling, 56 (15%) had a rating of 14 or more out of 20 for severity and frequency.

Table II.   Frequency of drooling by GMFCS level and topographical pattern
 Proportion of all CP (%)Study prevalence of drooling (%)
  1. GMFCS, Gross Motor Function Classification System.

GMFCS level I
 Hemiplegia19.312.5
 Diplegia11.14.6
 Quadriplegia2.750.0
GMFCS level II
 Hemiplegia12.037.3
 Diplegia12.218.12
 Quadriplegia4.460.0
GMFCS level III
 Hemiplegia1.666.7
 Diplegia5.125.0
 Quadriplegia4.563.0
GMFCS level IV
 Hemiplegia0.380.0
 Diplegia2.233.3
 Quadriplegia12.162.1
GMFCS level V
 Hemiplegia0.0
 Diplegia0.1100.0
 Quadriplegia12.480.0
Weighted prevalence (95% confidence interval)100.039.6 (33.8–45.4)

Factors associated with drooling

A comparison of the groups of children with and without a problem with drooling is shown in Table III. Sex and an additional diagnosis of an autism spectrum disorder were not associated with drooling, whereas epilepsy and intellectual disability were significantly associated with drooling on univariable analysis. Additional variables that were significantly associated with drooling beyond 4 years of age were special education, a non-spastic motor type, a quadriplegic or hemiplegic topographical pattern of motor impairment, inability to control head posture, GMFCS levels IV or V, limited or no useful speech, poor lip closure, anterior open bite, and eating difficulties.

Table III.   Comparison of children with and without a drooling problem
 No drooling (n=209)Drooling (n=153) p value
  1. GMFCS, Gross Motor Function Classification System.

Mean age, y (SD)11.0 (1.8)10.8 (1.7) 
Sex, n (%)0.266
 Male119 (55.3)96 (44.7) 
 Female90 (61.2)57 (38.8)
Type of schooling, n (%)<0.001
 Mainstream156 (76.5)48 (23.5) 
 Special and mainstream5 (31.2)11 (68.8)
 Special47 (34.1)91 (65.9)
 No formal schooling0 (0.0)2 (100.0)
Predominant motor type, n (%)0.014
 Spastic194 (60.4)127 (39.6) 
 Ataxic5 (31.3)11 (68.7)
 Dyskinetic1 (16.7)5 (83.3)
 Hypotonic9 (47.4)10 (52.6)
Topographical pattern, n (%)<0.001
 Hemiplegia86 (69.9)37 (30.1) 
 Diplegia82 (79.6)21 (20.4)
 Quadriplegia41 (30.2)95 (69.8)
Holds head up, n (%)<0.001
 All of the time192 (65.8)100 (34.2) 
 Some of the time14 (31.8)30 (68.2)
 None of the time1 (4.4)22 (95.6)
GMFCS, n (%)<0.001
 Level I65 (87.8)9 (12.2) 
 Level II84 (66.7)42 (33.3)
 Level III29 (53.7)25 (46.3)
 Level IV18 (41.9)25 (58.1)
 Level V12 (19.1)51 (80.9)
Comorbidities, n (%)
 Epilepsy39 (38.6)62 (61.4)<0.001
 Intellectual disability75 (45.2)91 (54.8)<0.001
 Autism spectrum disorder15 (55.6)12 (44.4)0.812
Speech, n (%)<0.001
 No problems with speech129 (89.0)16 (11.0) 
 Understood by strangers24 (58.5)17 (41.5)
 Difficult for strangers to understand28 (51.9)26 (48.1)
 Limited speech9 (36.0)16 (64.0)
 No useful speech16 (17.2)77 (82.8)
Lip closure, n (%)<0.001
 Can use a straw153 (80.1)38 (19.9) 
 Leakage occurs16 (21.3)59 (78.7)
 Mouth is always open18 (25.7)52 (74.3)
Dentition, n (%)<0.001
 Within normal limits130 (65.3)69 (34.7) 
 Anterior open bite28 (34.2)54 (65.8)
Eating, n (%)<0.001
 No difficulties169 (77.5)49 (22.5) 
 Difficulty with hard foods24 (39.3)37 (66.7)
 Food must be mashed/puréed12 (34.3)23 (65.7)
 Fully or partly tube fed1 (2.4)41 (97.6)

Discussion

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

This study makes an important contribution to the body of knowledge about drooling in children with CP by providing a population-based estimate of the prevalence of abnormal drooling in a cohort of children aged between 7 years and 14 years. Using the VCPR as our sampling frame enabled us to perform this research in a known population of children with CP, thereby allowing us to assess and adjust for participation bias. Our adjusted estimate of 40% across the full range of severities was very similar to the estimate of 37% reported in 211 children who attended or resided at the Belgian Institute for the Treatment of Cerebral Palsy,22 but higher than the estimate of 22% in children registered with the Northern Ireland Cerebral Palsy Register.20 The differences between the two population studies may reflect differences in the severity of drooling captured by each data collection method. Our study attempted to identify any degree of problem with drooling, whereas the Northern Ireland study specifically asked about excessive drooling.

The higher rates reported from some earlier studies most likely reflect the frequency of drooling in the most severely affected children. In our study, the prevalence of drooling for children with a quadriplegic distribution of motor impairment, and for those functioning in GMFCS levels III to V, was around 80%. This is similar to the figure of 78% estimated in a sample of 56 young people attending a special school in New York, USA.23 In comparison, frequencies of 58% and 65% were reported in studies of children attending special schools in the UK and India respectively.18,21 These higher estimates of the rates of drooling may largely be the result of the different sampling frames and measurement methods used. On the other hand, more consistency was seen for estimates of the frequency of severe drooling in CP, with 18% of Indian children reported to have a severe drooling problem,18 and 10% of Swedish children with CP presenting with embarrassing drooling.28 In our study 15% of parents reported that the frequency and severity of drooling rated at least 14 on a 20-point scale. In our experience, this is the level of severity at which parents might seek intrusive interventions.

Comparison of drooling frequencies between cohorts is not only affected by the clinical characteristics of the sample but also by the methodology adopted. Even in population cohorts, the age at which the presence of drooling is ascertained will affect the estimate because of maturational considerations, spontaneous remission, treatment effects, and changes in motor severity distributions as a consequence of early deaths. Hegde and Pani18 reported that 16% of their cohort had gained control of their drooling, either spontaneously or as a result of treatment, but it is unclear whether this percentage included children who drooled at a very young age when drooling is generally accepted as normal. Parkes et al.20 also reported a decline in speech problems and drooling over time. Our study assigned children to the drooling group if they had ever had a problem with drooling past the age of 4 years, and was intended to capture those whose problems had resolved. In fact, many families in the current study reported improvement over time, with 11 (7%) stating that their child no longer had a drooling problem. Four of these children had previous surgery for saliva control. The reasons given by the other seven families for the improvement were normal maturation in combination with speech therapy (n=2), Chinese acupressure (n=1), and reminders to swallow/wipe (n=1). Further research into changes in saliva control over time would, however, be useful to inform the most appropriate direction of clinical services.

Our large sample size allowed us to assess associations between drooling and several other demographic and clinical variables. The highest frequency of drooling was reported in children with severe involvement of all four limbs, no independent ambulation, poor oromotor function, and poor head control. Compared with children without a drooling problem, a higher proportion of children with non-spastic motor types drooled. This is somewhat in contrast to the findings reported by Hegde and Pani, where the highest prevalence of drooling was among children with a mixed motor type (60%) and lowest among children with athetosis (22%).18 Although numbers were small in our study, 83% of children with dyskinetic (athetoid or dystonic) CP were reported as having a drooling problem. This higher prevalence figure is consistent with research from the Netherlands, which provided some evidence that children with dyskinetic CP produce more saliva from their parotid and submandibular glands, presumably because of hyperkinetic oral motor activity.8 We also found a significant association between drooling and intellectual impairment, an association that has also been reported by other researchers.15,20 The clear finding of an association between intellectual disability and drooling from the two population samples indicates a possible need for further research to explore drooling among people with intellectual disability. Moreover, the presence of intellectual disability may affect treatment options and should be considered carefully when developing a treatment regime.

Our study found a relationship between limited speech and drooling, an association that has some support in the literature.14 In Senner et al.’s small study of 14 children with CP, the severity of drooling was significantly related to the severity of dysarthria. Parkes et al.20 did not specifically search for associations between oromotor impairment and drooling, but observed that motor speech impairment, moderate to severe communication impairment, problems with swallowing and chewing, and drooling were all more common in higher GMFCS levels. Overall, this reinforces the need for a multifactorial approach to drooling, and stresses the importance of focussing on improving saliva control in children with severe oromotor impairment, as excessive drooling can result in distressing social consequences for the child and family. Further investigation of the component factors that link specific oromotor skills to drooling would also be useful. Our results suggest that lip seal, a component of mouth closure, is a predictor of drooling, and this factor has been frequently targeted in the treatment of drooling.13,29 The relationship between lip seal and drooling has been reported in several studies.13,15,18,21 The results from Morales-Chavez’s descriptive study of 50 individuals with CP who attended a dental clinic for people with disabilities demonstrated that malocclusions were common, with 30% of their sample having anterior open bites and associated difficulties with lip seal and drooling.19

Validation and use of a drooling questionnaire in subsequent research would enable the study to be more easily replicated and facilitate comparisons across studies. Although the current study elicited a broad range of data about drooling in children with CP and highlighted key factors associated with it, more detailed analysis is still required of the functional components of the motor and sensory skills required for saliva control. In addition, future research is needed to investigate the prevalence problem of drooling in other cohorts, such as adults with CP, or individuals with specific acquired neurological impairments.

Conclusion

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

This study collected data representative of the population of children with CP living in Victoria, Australia, and the results provide a more reliable estimate of drooling prevalence than previously found in the published literature. Drooling was confirmed as a problem for 40% of 7- to 14-year-old children with all levels of motor severity. These findings highlight the magnitude of the problem of drooling in CP and provide a foundation for further investigation into targeted interventions for this under-researched and important problem.

Acknowledgements

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

We are grateful to Tess Lionti and Christine Westbury who contacted families and collected data for this study, and Janet Walstab who provided statistical advice. We acknowledge the assistance of the Marian & E.H. Flack Trust for providing funding for this project and Victorian Medical Insurance Agency Ltd, the name behind Professional Services Australia, and the Victorian Department of Health that funded the VCPR. The study was also supported by the Victorian Government’s Operational Infrastructure Support Program.

References

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References