Surgical treatment of velopharyngeal dysfunction: Incidence and associated factors in the Swedish cleft palate population ☆

Introduction: Speech in children with cleft palate can be affected by velophar - yngeal dysfunction, which persists after primary palate repair. The incidence of surgery to correct velopharyngeal dysfunction in this patient group has previously been reported as 2.6–37%. We aimed to investigate the incidence of velopharyngeal dysfunction surgery in Swedish children with cleft palate and to examine potential associations of independent variables with this incidence. Methods: In this cohort study, we analysed data from the Swedish cleft lip and palate quality registry for 1093 children with cleft palate with or without cleft lip. Kaplan–Meier analysis was used to estimate the risk of having velopharyngeal dysfunction surgery. Multivariable Cox proportional hazards models were used to estimate the associated effect of cleft subtype, additional diagnoses, gender, and age at and number of stages for primary palate repair on the primary outcome.

Cleft palate with or without cleft lip (CP ± L) requires surgical repair early in life to avoid negative impacts on feeding and speech. 1 The goal of palate repair is to separate the oral and nasal cavities and provide velopharyngeal closure, which is required for functional speech, eating, and drinking. 2Despite completed primary palate closure, some children with CP ± L have persisting velopharyngeal dysfunction (VPD).VPD is manifested by speech symptoms such as hypernasal resonance, audible nasal air leakage, and reduced pressure in high-pressure consonants.Some children actively try to compensate for VPD by changing articulatory place and/or manner for certain consonants. 3PD can have a negative impact on speech intelligibility 4 and, subsequently, health-related quality of life. 5,6VPD is usually treated surgically with the aim of creating a functional seal between the oral and nasal cavities during speech. 7Previous reports of the incidence of VPD surgery in patients with CP ± L are highly variable, between 2.6% and 37%, [8][9][10][11] with inconsistencies as to which surgical procedures are counted.
CP ± L is divided into cleft subtypes depending on the extent of the cleft.The most common subtypes involving the palate are cleft soft palate (SP), cleft hard and soft palate (SHP), unilateral cleft lip with cleft hard and soft palate (UCLP), and bilateral cleft lip with cleft hard and soft palate (BCLP).Previous studies have reported correlations between the incidence of VPD surgery and cleft subtype, 9,[12][13][14] cleft width, 15 gender, 13,16 syndromes or additional malformations, 8,17 and chronic otitis media with effusion. 17There have also been reports of no such significant correlations. 8,129][20][21][22] A recent systematic review found the lowest level of VPD after primary palatoplasty at the age of 6-12 months, 23 while Skolnick et al. found a higher incidence of VPD surgery after primary palate repair before the age of 10 months 24 and Klintö et al. reported negative speech results after complete primary palate repair after the age of 25 months. 25Higher incidences of VPD following primary palate repair after 13 months 26 and after 18 months 27 were described in two studies with less stringent methodologies.
In Sweden, around 175 individuals are diagnosed with cleft lip with or without cleft palate each year. 25They are treated at one of six regional cleft lip and palate (CLP) centres from birth until the age of 19 years.The treatment centres used varying surgical protocols in 2009-2022.Centre 1 performed palatoplasty in one stage for SP and SHP and in two stages for UCLP and BCLP.Centres 2 and 4 used two stages, Centres 5 and 6 used one stage, and Centre 3 transitioned from one to two stages during the study period. 25Surgical methods, number of chief operators, and annual caseloads at each centre have previously been described by Klintö et al. 25 All six centres are connected to the Swedish quality registry for patients born with CLP (CLP Registry).Baseline data recorded at the first visit includes cleft type, gender, additional diagnoses, and whether the child was born in Sweden or abroad. 28Data for children born abroad regarding surgery performed in their country of birth often lacks validity.Each treatment centre continuously registers surgical procedures, treatment outcomes for babbling/early speech, and speech and dentofacial development at specific time points. 28Since 2009, all children born in or who moved to Sweden are included.In 2020, the average coverage degree of the CLP Registry for children born in 2009-2018 was 95%, and cleftrelated surgeries had an average reporting degree of 92%. 28he aim of this study was to estimate the incidence of VPD surgery in children born with CP ± L in Sweden and examine whether the risk of VPD surgery was associated with cleft subtype, gender, syndromes or additional malformations, age at primary soft palate repair, age at complete primary palate repair, and number of stages for primary palate closure.

Study design
This register-based cohort study was based on data retrieved from the CLP Registry on 11 April 2022 via the Centre of Registries South, Lund, Sweden.The data had been registered in baseline, surgery, and speech forms for children born between 2009 and 2017 with CP ± L (1367 children).All data was manually examined, and potential faulty registrations were checked against medical records.This resulted in corrections for 71 participants.In addition, 55 surgical procedures which had been mistakenly registered as 'primary' were altered to 'secondary' (fistula repairs, re-repairs, and velopharyngeal flaps, which are only performed as secondary procedures, and palate repairs Journal of Plastic, Reconstructive & Aesthetic Surgery 90 (2024) 240-248 performed several years after primary palate repair).This study adheres to the STROBE guidelines.

Study sample
Of the 1367 children in the registry, 25 children were registered as deceased or had moved abroad and were therefore excluded.Other reasons for exclusion were submucous clefts or clefts that, after validation, were found not to involve the palate (10), cleft surgery abroad (52), no registrations apart from baseline, indicating medical reasons for missing primary palate surgery (13) and not born in Sweden (174).This resulted in a total of 1093 participants, for which characteristics are presented in Table 1.
During the study period, palate repair was usually performed either in one stage at the age of 9-18 months or in two stages, with soft palate closure at the age of 6 months and hard palate closure around 2 years. 25or all children, age at primary soft palate repair was divided into four categories based on quartile values: < 7 months, 7-9 months, 9-13 months, and > 13 months.Age at complete primary palate repair for all children was divided into five categories based on results from previous studies [24][25][26][27] : < 10 months, 10-13 months, 13-18 months, 18-25 months, and > 25 months.

Outcome measure
VPD surgery was defined as palatal and/or velopharyngeal surgical procedures registered as 'secondary'.In Sweden, a videofluoroscopic and/or nasoendoscopic evaluation of the velopharyngeal function during speech is usually a necessary precedent for VPD surgery.These evaluations are rarely performed on children younger than 3 years of age since it may be difficult for them to participate in the assessment; therefore, only surgical procedures performed from the age of 3 years and upwards were included as VPD surgery in this study.'Re-repair of palate' and 'secondary soft/hard/hard and soft palate repair' were all categorised as 're-repair' as these procedure codes were used synonymously in the CLP Registry.More than one procedure registered on one occasion was categorised as a 'combination of procedures'.Included procedures are presented in Table 2.

Statistical analysis
Patient characteristics were summarised by descriptive statistics.Patients were followed from birth until the first VPD surgery or study end (11 April 2022), where patients who had not experienced a VPD surgery were censored.
Kaplan-Meier analysis was used to estimate the risk for a child born with CP ± L of having VPD surgery.Univariable and multivariable Cox proportional hazards models, including the independent variables cleft subtype, gender, age at primary soft palate repair, age at complete primary palate repair, and number of primary palate surgeries, were used to estimate the associated effect of each variable on the risk of having VPD surgery, adjusting for all other variables in the model.Children with a diagnosis of Robin sequence (RS), syndromes, or other malformations were classified as CLP+ and the remaining children as CLP−, and  the statistical analysis described above was performed for this independent variable as well.
We expected an association between age at primary soft palate repair, age at complete primary palate repair, and number of primary palate surgeries.These associations were evaluated graphically, and separate multivariable Cox regression models were, therefore, fit for these variables while stratifying for gender and adjusting for cleft subtype and CLP ± .
The statistical significance level was set at p < 0.05.Statistical analysis was performed using IBM SPSS Statistics version 28.0.1.1.

Incidence of VPD surgery
The mean follow-up time from birth was 8.9 years (SD 2.6).During a total follow-up time of 8836 years, 200 of the children received VPD surgery (event, Table 1), and 893 did not.41 children had more than one occurrence of VPD surgery.
Using Kaplan-Meier analysis, the risk of having VPD surgery before the age of 13 years was estimated to be 25.6%.The same analysis excluding fistula repairs gave an estimated risk of 22.0%.

Univariable regression analysis
A comparison of Kaplan-Meier curves stratified by the independent variables previously described indicated no major deviations from the proportional hazards assumption for any of the independent variables except for gender.The effect of gender was found to vary over time, as VPD surgery was more common for girls than boys before 6 years of age but more common for boys than girls after the age of 6 years (Table 3, Figure 1).
Univariable Cox regression models demonstrated a significantly lower risk of having VPD surgery for SP compared to SHP (HR = 0.36, p < 0.001) for females compared to males after the age of 6 years (HR = 0.61, p = 0.037), for primary soft palate repair after the age of 13 months compared to before the age of 7 months (HR = 0.59, p = 0.017), and for complete primary palate repair before the age of 10 months (HR = 0.45, p < 0.001), at 10-13 months ( HR = 0.55, p = 0.005) and 13-18 months (HR = 0.35, p < 0.001) compared to the age of 18-25 months.A significantly higher risk was found for primary palate repair in two stages (HR = 2.29, p < 0.001) or three stages (HR = 5.16, p < 0.001) compared to one stage (Table 3).

Multivariable regression analysis
The variables age at complete primary palate repair, age at primary soft palate repair, and number of stages for primary palate repair were associated (Figure 2).To avoid multicollinearity, separate multivariable Cox regression models were fit for each of these variables, all adjusted for gender (by stratification), cleft subtype, and CLP ± .
The risk of VPD surgery increased with an increasing number of stages of primary palate repair (Table 3, multivariable model 1: HR = 2, p < 0.001, for two stages and HR = 4.48, p < 0.001 for three stages vs. one stage).A lower age at complete primary palate repair was associated with a reduced risk of VPD surgery (Table 3, multivariable model 2: HR = 0.58, p = 0.013 for 10-13 months, and HR = 0.36, p < 0.001 for 13-18 months, compared to > 18 months).Primary SP repair after 13 months of age was associated with a lower risk than repair before 7 months of age (Table 3, multivariable model 3: HR = 0.53, p = 0.004).
In a multivariable Cox regression model including both age at complete primary palate repair and number of stages for primary palate repair, adjusted for gender, cleft subtype, and CLP ± , only the HR for repair in three stages compared to one stage remained statistically significant (Table 3, multivariable model 4: HR = 3.26, p = 0.005).
Finally, in a multivariable Cox regression model including both age at primary soft palate repair and number of stages for primary palate repair, adjusted for gender, cleft subtype, and CLP ± , the risk of VPD surgery increased with an increasing number of stages of primary palate repair (Table 3, multivariable model 5: HR = 1.97, p < 0.001 for two stages and HR = 4.50, p < 0.001 for three stages vs one stage).However, there was no statistically significant association between the risk of VPD surgery and age at primary SP repair in this model.

Discussion
The aim of this study was to investigate the incidence of VPD surgery in Swedish children with CP ± L and to examine potential associations of independent variables with this incidence.

Incidence of VPD surgery
The incidence of VPD surgery was 25.6%.][10][11] The referenced studies varied in follow-up time and definitions of VPD surgery, which might explain some of the variation in incidence, but there do not appear to be any clear connections between similar research methodologies and incidence rates.The variations might reflect variations in VPD incidence between populations, but the proneness to perform VPD surgery could also differ between institutions and surgeons. 30

Association with timing and staging of primary palate repair
Previous studies have provided evidence both for and against significant associations between age at primary palate repair or number of palate surgeries and VPD. 18,23,25,31,32The same is true for number of primary palate surgeries. 20,21In our sample, complete primary palate repair after the age of 18 months and in more than one stage were both associated with a higher risk of VPD surgery when analysed independently of each other, but when analysed in the same model, only the increased risk associated with repair in three stages compared to one stage     remained statistically significant.As seen in Figure 2, the mean age at complete palate repair was higher for the two-and three-stage groups than the one stage group.Given the correlation between number of primary surgeries and age at primary surgery, it could not be determined which of the two variables was the potential cause for the increased risk.Further analysis of groups with fewer variations in treatment factors would eliminate potential confounders but would also decrease sample sizes.

Association with cleft subtype
In agreement with previous studies, 12,33,34 we found a significantly lower risk of VPD surgery for patients with SP compared to other cleft subtypes but no significant difference in risk between other cleft subtypes.Cleft subtype influenced the timing of complete primary palate repair especially in the cases of two-stage repair, as clefts of the soft palate were repaired in one stage by all centres and thus completely repaired earlier than clefts of both the hard and SP.

Association with additional diagnoses
Around 30% of children born with CLP have additional malformations. 35Some studies have found an association between VPD and syndromes, RS, and additional malformations, 8,17 whereas other studies found no such association. 24,36,37In the  present study, no difference in the risk of VPD surgery was found for the CLP+ group compared to the CLP− group.A detailed analysis of the separate subgroups within this category is warranted but is not within the scope of this study.

Velopharyngeal function
The incidence of VPD surgery is influenced not only by the level of speech impairment but also by the inclination to operate at each treatment centre and the wishes of the patient and family. 38Therefore, the incidence of VPD surgery is not a sufficient outcome measure of success of primary palate repair as it may not be synonymous with the incidence of VPD. 30 In our next study, we plan to map velopharyngeal function based on perceptual speech evaluations recorded in the CLP Registry and to examine the correlation between the incidence of VPD surgery and VPD.

Strengths and limitations
This was a longitudinal study of a near-complete national patient cohort.The coverage and reporting degrees for surgical procedures were high and corresponded to typical coherence in prospective randomised multicentre studies where a drop-out level of around 10% is generally expected. 31,32,39This minimised the risk of selection bias.However, using retrospective data allows for potential errors or differences in registration of data and evaluation of outcomes.This must be taken into consideration when interpreting the results.Using clinical registry data in research restricts the choice of variables to those included in the specific registry.In this case, the lack of data on number of surgeons, surgical technique, cleft width, as well as possible unidentified confounders was a limitation of the study.Conclusions regarding causality cannot be drawn here; the results should be interpreted as suggestions of associations between variables.
Reasons for performing surgery were not included in the CLP Registry.Associations between palatal fistulas and VPD have previously been found, 8,40 but fistulas can be repaired to correct nasal regurgitation of food and drink as well as VPD.We decided to include fistula repairs from 3 years and upwards as VPD surgery; however, some of the procedures interpreted as VPD surgery in this study may in fact be secondary palate surgeries for other reasons than VPD.
Surgical protocols for repairing clefts involving both the hard and soft palate differed between the Swedish CLP treatment centres.Number of primary palate surgeries and age at primary soft and complete palate repair were therefore, to a high degree, dependent on the treatment centre.Treatment centres could, therefore, not be adjusted for in the regression models, and unknown factors associated with treatment centres, such as surgical skill, surgical techniques, health care infrastructure, and differences in patient populations, might have influenced the difference in outcomes between the centres. 25There may also have been other sources of unmeasured confounding.We cannot rule out that the age at, and staging of, primary palate surgery was related to patient characteristics, which were not measured in our study (e.g., caregivers' preferences, unregistered medical conditions, and patient growth patterns).

Conclusions
The risk of having VPD surgery before the age of 13 years for Swedish children born with CP ± L was 25.6%.This is in agreement with previous studies.Complete primary palate repair after the age of 18 months and in more than one stage was associated with a significantly higher risk of having VPD surgery, but additional unmeasured factors may also have affected this outcome.

Statement of ethical approval
Ethical approval was obtained from the Swedish Ethical Review Authority (reference no.2020-00227).Data access was approved by the consultation group for quality registries, care databases, and preparation in the Skåne Region (reference no.036-22).

Figure 1
Figure 1 Kaplan-Meier curve showing probability of no velopharyngeal dysfunction (VPD) surgery after 3 years of age for children born with CP ± L in Sweden, stratified by gender.

Figure 2
Figure 2Boxplots of age at primary soft palate repair and age at complete primary palate repair stratified by number of primary palate surgeries.

Table 1
Study sample characteristics.
SP, cleft soft palate; SHP, cleft hard and soft palate; BCLP, bilateral cleft lip with cleft hard and soft palate; UCLP, unilateral cleft lip with cleft hard and soft palate; RS, Robin sequence.

Table 2
Types, numbers, and proportions of velopharyngeal dysfunction surgery procedures.

Table 3
Results of Cox regression analysis modelling associations of independent variables to risk of having velopharyngeal dysfunction surgery.
Multivariable model 1 includes number of primary palate surgeries, model 2 includes age at complete primary palate repair, model 3 includes age at primary soft palate repair, model 4 includes age at complete primary palate repair and number of primary palate surgeries, and model 5 includes age at primary soft palate repair and number of primary palate surgeries.SP, cleft soft palate; SHP, cleft hard and soft palate; BCLP, bilateral cleft lip with cleft hard and soft palate; UCLP, unilateral cleft lip with cleft hard and soft palate; RS, Robin sequence.