Speech and Phonology in Swedish-Speaking 3-Year-Olds with Unilateral Complete Cleft Lip and Palate following Different Methods for Primary Palatal Surgery

Abstract

Objective

To describe and compare speech and phonology at age 3 years in children born with unilateral complete cleft lip and palate treated with three different methods for primary palatal surgery.

Design

Prospective study.

Setting

Primary care university hospitals.

Participants

Twenty-eight Swedish-speaking children born with nonsyndromic unilateral complete cleft lip and palate.

Interventions

Three methods for primary palatal surgery: two-stage closure with soft palate closure between 3.4 and 6.4 months and hard palate closure at mean age 12.3 months (n = 9) or 36.2 months (n = 9) or one-stage closure at mean age 13.6 months (n = 10).

Main Outcome Measures

Based on independent judgments performed by two speech-language pathologists from standardized video recordings: percent correct consonants adjusted for age, percent active cleft speech characteristics, total number of phonological processes, number of different phonological processes, hypernasality, and audible nasal air leakage. The hard palate was unrepaired in nine of the children treated with two-stage closure.

Results

The group treated with one-stage closure showed significantly better results than the group with an unoperated hard palate regarding percent active cleft speech characteristics and total number of phonological processes.

Conclusions

Early primary palatal surgery in one or two stages did not result in any significant differences in speech production at age 3 years. However, children with an unoperated hard palate had significantly poorer speech and phonology than peers who had been treated with one-stage palatal closure at about 13 months of age.

Keywords

phonology primary palatal surgery speech unilateral cleft lip and palate

Successful primary palatal surgery is crucial for speech development in children born with cleft palate (CP). Despite decades of research on speech results after primary palatal surgery, evidence about which methods give the best speech results is still missing. Today, the chosen method for primary palatal surgery in children born with CP seems to depend more on the surgeon's previous experience and training from mentors than on any evidence of which method provides the best result (Friede, 2009). Several general factors complicate the evaluation of intervention in CP, such as multidimensionality of outcome, duration of follow-up, reproducibility and validity of outcome measures, diversity of management, and small sample size (Roberts et al., 1991). The possibility of comparing different treatment procedures has also been limited by a lack of standardized methods for collection and analysis of speech data (Lohmander and Olsson, 2004; Sell, 2005). However, in recent decades greater awareness has developed regarding the necessity to produce better evidence, meaning that interest in evaluating standardized surgical procedures in a prospective way has increased.

In choosing a surgical approach, many different aspects must be considered. Methods of primary palatal repair may vary regarding timing, staging, sequence, and techniques. Surgical caseload and skill may also affect speech outcome (Williams et al., 1999). Surgery that results in maxillary scarring may have a negative impact on facial growth (Sommerlad, 2006). Some researchers argue that facial growth benefits from delayed palatal closure (Friede, 2007), whereas speech development benefits from palatal closure as early as possible (Peterson-Falzone, 1996).

To date, the basis for a decision on the most effective method for primary palatal repair is rather weak. In a recent review of articles on speech outcome after primary palatal surgery based on assessment from recordings, the most common surgical procedures were Wardill-Kilner push-back closure, the Van Langenbeck technique, and a two-stage procedure with delayed hard palate closure (Lohmander, 2011). The review revealed large variations regarding time from surgery to follow-up and the age range of included patients. Also, in half of the 34 studies reviewed, the studied groups were relatively small (fewer than 50 participants). In summary, no significant differences have been reported in speech outcomes in adults after one-stage surgery with the Van Langenbeck technique and Wardill-Kilner push-back closure (Pigott et al., 2002; Farzaneh et al., 2008). When speech outcome was compared after the Furlow procedure and after the Van Langenbeck technique, contradicting results were reported (Spauwen et al., 1992; Van Lierde et al., 2004). Speech outcome after two-stage procedures has been reported to be at least as good as or even better than speech outcome after one-stage procedures (e.g., Van Demark et al., 1989; Lohmander et al., 2006; Lohmander et al., 2012). For example, the speech results in a study of children treated with a two-stage procedure (Persson et al., 2002) were better than those in a study of children treated with a one-stage procedure with minimal incision technique (Nyberg et al., 2010). Both studies reported speech outcome data for 5-year-old children born with isolated CP. Very similar methodologies were used for the assessments in those two studies. In other studies, the methods for speech assessment have varied greatly, making comparisons difficult. Therefore, issues such as surgical timing, techniques, and staging, as well as surgical caseload and skill, cannot be compared reliably in terms of their impact on speech production.

In recent studies of young children born with cleft lip and palate (CLP), approximately 50% of the children display impaired speech/phonology around age 3 years, regardless of whether the palate was repaired in one or two stages, timing, and technique used for primary palatal repair (e.g., Morris and Ozanne, 2003; Chapman et al., 2008; Lohmander and Persson, 2008; Willadsen, 2012). Compared to age-matched typically developing children, significantly lower percentages of correct consonants were found for children who received two-stage primary palatal repair (Lohmander and Persson, 2008; Klintö et al., 2013) and for children who underwent one-stage repair (Scherer et al., 2008). Willadsen (2012) investigated the influence of timing of hard palate closure on early speech development in children born with unilateral CLP (UCLP) treated with a two-stage palatal repair. The author found that children with an open hard palate at 3 years of age had poorer phonology than children who had the hard palate closed at 1 year of age. Chapman et al. (2008) were interested in the impact of age and lexical status at the time of primary palatal surgery on speech outcomes of preschoolers (median age 39 months; range, 33 to 42 months) with cleft palate with or without cleft lip (CP±L). Their results indicated better articulation and resonance outcomes in children who were less lexically advanced at the time of palatal surgery (at mean age 11 months; range, 7 to 14 months) than in children who were more lexically advanced at time for palatal surgery (at mean age 15 months; range, 12 to 23 months). However, varying cleft types, age ranges, and surgical methods in their study may have influenced the results.

Phonological disorders in children born with CP may also have other etiologies. Morris and Ozanne (2003) studied expressive language in 3-year-olds with CP of varying types who had their CP repaired between 6 and 12 months of age. They found a subgroup (about half of the children) of children at risk for continued language and speech impairment. They discussed possible etiologies, such as structural/anatomical deficits, cognitive/linguistic delays, and language/phonological disorders. In the interpretation of results for phonology and other expressive language in children born with CP, it is important to keep in mind that not only the CP and its treatment, but also other underlying cognitive and linguistic variables, may influence a child's development.

It should also be noted that the incidence of otitis media with effusion and related mild to moderate hearing loss is high among children born with CP (Aniansson et al., 2002; Flynn et al., 2009). The associated hearing loss decreases a child's access to speech (Roberts et al., 2004). Lohmander et al. (2011) found a significant negative correlation between mild hearing impairment and consonant inventory at age 12 months; however, this correlation was no longer significant at 18 months of age. Hearing status needs to be considered when assessing speech and language in children born with CP.

Because the impact of surgery on the development of speech and phonology is unclear, there is a need for investigations of speech outcome after different surgical procedures for primary palatal repair. The purpose of the present study was to describe and compare speech and phonology in children with complete UCLP treated with three different surgical methods. The research question was whether there would be, at that age, a difference related to the different procedures for primary palatal repair in terms of (1) percent correct consonants adjusted for age (PCC-A), (2) percent active cleft speech characteristics, (3) occurrence of phonological processes, and (4) audible nasal resonance and nasal air leakage.

Methods

Participants

A total of 30 children with complete UCLP were included in the study. All were monolingual speakers of Swedish. Children with known additional malformations and/or syndromes were excluded. Twenty of the children formed a consecutive group, born between 1997 and 2003, and all were patients at the Sahlgrenska University Hospital, Gothenburg, Sweden. The children were participants in the Scandcleft project (Semb, 2001). They were randomized for hard palate closure at 12 or 36 months of age, following early soft palate repair together with lip closure at a mean age of 4.6 months (range, 3.4 to 6.4 months). The first group, two-stage closed (TSC), was randomized for early hard palate closure and underwent surgery at a mean age of 12.3 months (range, 11.6 to 13.7 months). The second group, two-stage open (TSO), was randomized for late hard palate closure and underwent surgery at a mean age of 36.2 months (range, 35.6 to 37.1). Two fistulas were reported in the TSC group and none were seen in the TSO group. No other postoperative complications were reported in either two-stage group.

The third group, one-stage closed (OSC), was a consecutive series of 10 children, born between 2005 and 2008, who were patients at Skåne University Hospital, Malmö, Sweden. They were treated with one-stage palatal closure at a mean age of 13.6 months (range, 11 to 15 months). Of these 10 children, six had previously undergone a primary lip repair at a mean age of 4.5 months (range, 4 to 6 months), and four had undergone lip adhesion at mean age 3.75 months (range, 3 to 6 months) followed by primary lip repair at a mean age of 9 months (range, 6 to 11 months). No fistulas or other postoperative complications were reported in the OSC group.

Of the children treated at Sahlgrenska University Hospital (the TSC and TSO groups), two children were excluded; one was unable to cooperate and one recording was missing for the other child. The resulting participant group consisted of 28 children: 9 in the TSC group, 9 in the TSO group, and 10 in the OSC group.

None of the children in the three groups received speech therapy from a speech-language pathologist. However, during their first 3 years in life, all of them met with a speech-language pathologist five to seven times for follow-up and counseling.

Ethical Approval

The participation of the cleft center in Gothenburg in the Scandcleft project was approved by the Regional Ethical Review Board of Gothenburg (R257-97). Enrollment of the patients treated in Malmö was approved by the Regional Ethical Review Board of Lund (no. 548/2008). All parents had given written informed consent for participation.

Methods for Soft and Hard Palatal Closure

The 18 children treated at Sahlgrenska University Hospital underwent early soft palate repair with delayed hard palate closure at the age of 1 or 3 years (TSC or TSO) according to the Scandcleft protocol. All surgeries were performed between the years 1997 and 2004 by two surgeons. The soft palate repair was performed with a new procedure (Friede et al., 2013). The Scandcleft soft palate repair technique starts with a zigzag incision at the border between the soft and hard palate and continues anteriorly on the palatal shelf. This enables a small mucoperiosteal flap to be raised after incising along the cleft at the border between the nasal and oral mucosa all the way to the uvula. This mucoperiosteal flap is included in the oral layer, which is dissected from the muscles. A posteriorly and cranially based vomer flap is raised from behind the premaxillary suture and turned over to be sutured into the anterior half of the nasal layer, which facilitates the closure and anchors the soft palate to the vomer. The nasal layer is not released from the posterior part of the hard palate. For intravelar veloplasty, the muscles are dissected from the nasal mucosa without release of the deeper part of the tensor muscle, repositioned posteriorly, and sutured in the midline. The oral flaps are sutured along the midline, and the small mucoperiosteal flaps are used to cover the whole raw surface of the vomer flap. The cleft in the hard palate is left open for later closure in one layer by the use of a cranially based vomer flap. This procedure was used for all but one child, for whom a two-layer closure with mucoperiosteal flaps was used.

The 10 children treated at Skåne University Hospital underwent intravelar veloplasty, according to the method suggested by Sommerlad (2006), which was performed by one surgeon between the years 2005 and 2009. In four cases, however, lateral incisions were performed to prevent fistula formation due to tension.

Audiometry

Hearing at 3 years was assessed via audiometry by a pediatric audiologist, on the same day as the child's speech and language were assessed (Table 1). In the TSC group, three children were found to have mild hearing loss (21 to 40 dB hearing threshold level). Hearing data were missing for one child. Five children in the TSO group and three in the OSC group had mild hearing loss. The remaining children had normal hearing.

Table 1.

Number of Children with Hearing Impairment (21 to 40 dB HTL)^*

HI	TSC	TSO	OSC
No	5	4	7
Unilateral	2	4	2
Bilateral	1	1	1
Missing	1	0	0

HTL = hearing threshold level; HI = hearing impairment; TSC = two-stage repair with closed hard palate; TSO = two-stage repair with open hard palate; OSC = one-stage repair with closed palate.

Recording Procedure

All children were video recorded as part of the clinical routine at 3 years of age (TSC and TSO mean age 36 months, range 35.5 to 36.5 months; OSC mean age 36.9 months, range 35 to 39 months). The speech was recorded with the hard palate unrepaired in the TSO group. The recordings took place in a room at Sahlgrenska University Hospital or at Skåne University Hospital during interaction with a speech-language pathologist and lasted between 30 and 45 minutes. A video camera (Sony DCR-TRV30E) with an external microphone (Sony ECM-MS957) was used for the TSC and TSO groups. For the OSC group, a video camera (Canon HF10) with an external microphone (Sony ECM-M5957) was used. The microphone was placed in the middle of the table in front of the child.

A word-naming test was used. It had been developed in the Scandcleft Project to assess CP speech (Lohmander et al., 2009). The test consists of 32 pictures with the purpose of eliciting single words. An attempt by the child to produce the target word was counted as an elicited word, even if the production was unintelligible. If a child could not find the target word and semantic prompting failed, the speech-language pathologist said the target word and asked the child to imitate it. If the child refused to name the picture or said another word, the word was counted as not elicited. The median number of words elicited for each child was 31 (range, 17 to 32). The median number of words elicited with repetition was 11 (range, 1 to 20).

Editing of the Video Recordings

Sequences with word-naming of pictures as well as conversational speech samples were available for analysis. The single words in the word-naming test were chosen for analysis since they represented the best production of the child (Klintö et al., 2011) and to standardize the samples as much as possible (Klintö et al., 2013). For many of the children, most of the conversational speech was unintelligible and would have been excluded from the analysis. Therefore, it was not considered possible to get a representative and standardized speech sample with connected speech. The scale rating of nasality and audible nasal air leakage was performed as an overall assessment of all words in the sample. This is similar to the final method used in the Scandcleft project, in which hypernasality was assessed on single words in a word string (Persson et al., 2013).

The video recordings were edited in Corel Video-Studio Pro X4 (Corel Corporation). The child's production of the target word was followed by the test leader's repetition of the word. If a child could not produce the word on his or her own, the test leader's production of the target word was followed by the child's repetition. For each child, all single word productions were compiled into one digital video file. The files were presented to the judges in random order.

Phonetic Transcription and Rating of Passive Cleft Speech Characteristics

Phonetic transcription and rating of passive cleft speech characteristics of all video recordings was performed by the first author (main judge), who had more than 8 years' experience in the area of CP speech. A second independent external judge, with more than 19 years' experience with CP speech, retranscribed and re-rated all recordings for the calculation of intertranscriber and interrater agreement. One month later, the samples from all children were retranscribed and re-rated by the main judge, for calculation of intratranscriber and intrarater agreement. Narrow transcription based on the conventions of the International Phonetic Association (IPA) and Extensions to the IPA was used (International Phonetic Association, 1999). Recordings were played back as often as necessary. Hypernasality, hyponasality, and audible nasal air leakage were each rated on a scale from 0 to 3, on which 0 indicates normal resonance/no audible nasal air leakage and 3 indicates severe deviation of resonance/audible nasal air leakage occurring always or almost always. A computer with Windows Media Player (Microsoft) and headphones (Denon AH-D1001 or Sennheiser HD 218) was used. Prior to transcribing and rating, the judges had undertaken 13 hours of joint transcription and rating of video recordings of 3-year-old children born with CP who were not included in the study. The use of different phonetic symbols and diacritics and the grading of the four-point scales were discussed.

Analysis of Transcriptions

PCC-A values (Klintö et al., 2011, 2013) were calculated. Varying types of lisps and r-distortions were accepted as age adequate and scored as correct. Percent active cleft speech characteristics were calculated for each child (Klintö et al., 2011). The total number of phonological processes and the number of different types of phonological processes were assessed for each child. In addition, a qualitative analysis of phonological processes and CP-related processes was carried out (Klintö et al., 2013). Processes that occurred in at least 20% of all possible occurrences were judged as consistent, and those occurring less frequently were judged as sporadic (McReynolds and Elbert, 1981).

Agreement

Intratranscriber and intertranscriber agreement of consonant transcriptions for each sample of all children was calculated by means of percent agreement, point by point. The compared consonants had to be identically transcribed for place, manner, and voicing to be considered as agreed upon (Klintö et al., 2011). Mean intratranscriber agreement was 90% (range, 72% to 99%), and mean intertranscriber agreement was 77% (range, 48% to 90%).

Exact mean intra- and interrater agreement values, with and without an acceptance of one scale value difference on the four-point scale, were calculated. Levels of 80% for exact agreement and 90% when one scale difference was accepted were applied (Zarcone et al., 1991). Exact mean intrarater agreement was 79% for hypernasality and 86% for hyponasality and audible nasal air leakage. Exact mean interrater agreement was 43% for hypernasality, 46% for audible nasal air leakage, and 75% for hyponasality. With an acceptance of one scale value difference on the four-point scale, the mean intrarater agreement was 100% for all variables, and the mean interrater agreement improved to 71% for hypernasality, 100% for hyponasality, and 96% for audible nasal air leakage.

Further calculations on agreement were performed, since interrater agreement for hypernasality on the four-point scale with acceptance of one scale value difference was still regarded as poor. There was a small distance between the first two scale values for each speech variable, and these values were pooled into one point in the analysis. The variables and rating scales after pooling are presented in Table 2. After pooling, the exact mean intrarater agreement was 86% for hypernasality and 100% for both hyponasality and audible nasal air leakage. The exact mean interrater agreement after pooling was 64% for hypernasality, 100% for hyponasality, and 89% for audible nasal air leakage. With an acceptance of one scale value difference after pooling, the mean intrarater agreement was 100% for all variables. The mean interrater agreement percentages after pooling with an acceptance of one scale value difference were 100% for hyponasality and audible nasal air leakage and 75% for hypernasality.

Table 2

Description of Scale Values After Pooling of the Two First Scale Values Used in the Analysis of the Assessed Passive Cleft Speech Characteristics

Scale Value	Hypernasality and Hyponasality	Audible Nasal Air Leakage
0	Normal resonance or mild deviation	No occurrence or single/some occurrences
1	Moderate deviation	Frequently occurring
2	Severe deviation	Occurring always

Statistical Analysis

For all statistical analyses, P < .05 (two-tailed) was considered to indicate significant differences. Nonparametric statistics were used because of the small groups and skewed distribution of data. Descriptive statistics were presented with median values and ranges for the three groups: TSC, TSO, and OSC. The Kruskal-Wallis test was used for comparisons among group median values. For variables that displayed significant differences with the Kruskal-Wallis test, the Mann-Whitney U test with Bonferroni correction (P < .05 ≤ .017) was used for post hoc pairwise analysis of differences between group median values.

Results

PCC-A, percent active cleft speech characteristics, total number of phonological processes, and number of different types of phonological processes for each group, as well as group comparisons, are presented in Table 3. Considering the median values for all four outcomes, the results indicated best performance in the OSC group and worst in the TSO group. The median PCC-A values were 76% in the TSC group, 59% in the TSO group, and 86% in the OSC group. The median number of active cleft speech characteristics was 16 in the TSC group, 29 in the TSO group, and 6 in the OSC group. The median total number of phonological processes was 21 in the TSC group, 35 in the TSO group, and 13 in the OSC group. Finally, the median number of different types of phonological processes was eight in the TSC group, nine in the TSO group, and six in the OSC group. Significant differences, when tested with the Kruskal-Wallis test, were seen among the three groups for the outcomes percent active cleft speech characteristics and total number of phonological processes. Post hoc pairwise group comparisons revealed significantly better performance for the OSC group than for the TSO group in terms of percent active cleft speech characteristics and total number of phonological processes (Table 4).

Table 3

Comparisons Among Articulatory and/or Phonological Outcomes at 3 Years of Age (Kruskal-Wallis Test)

Outcomes	TSC^*, Median (Range) (n = 9)	TSO, Median (Range) (n = 9)	OSC, Median (Range) (n = 10)	χ²	P
PCC-A	76 (14–91)	59 (3–92)	86 (41–97)	5.641	.060
CSC	16 (3–51)	29 (4–64)	6 (0–26)	8.334	.015^†
NPP	21 (6–72)	35 (9–76)	13 (4–44)	6.821	.033^†
DPP	8 (3–12)	9 (4–16)	6 (3–13)	3.382	.184

TSC = two-stage repair with closed hard palate; TSO = two-stage repair with open hard palate; OSC = one-stage repair with closed palate; PCC-A = percent correct consonants adjusted for age; CSC = percent active cleft speech characteristics; NPP = total number of phonological processes; DPP = number of different types of phonological processes.

†

P < .05.

Table 4

Differences Between Groups when Analyzed Pairwise (Mann-Whitney U Test)

	Measures
	CSC^*		NPP		Hypernasality
Groups	z	P	z	P	z	P
TSC (n = 9) × TSO (n = 9)	–.662	.508	–.928	.353	–1.738	.082
TSC (n = 9) × OSC (n = 10)	–1.879	.060	–1.756	.079	–1.089	.276
TSO (n = 9) × OSC (n = 10)	–2.859	.004^†	–2.452	.014^†	–2.408	.016^†

TSC = two-stage repair with closed hard palate; TSO = two-stage repair with open hard palate; OSC = one-stage repair with closed palate; CSC = percent active cleft speech characteristics; NPP = total number of phonological processes.

†

P < .05 ≤ .017 with Bonferroni correction.

Qualitative analysis showed that the median number of consistent processes was 2 in the TSC group, 4 in the TSO group, and 1.5 in the OSC group. The median number of sporadic processes was 4 in the TSC group, 6 in the TSO group, and 4.5 in the OSC group. The most common phonological/articulatory processes used consistently are presented in Table 5. For explanation of the processes, see Appendix 1. Backing to palatal/velar/uvular was the most common process in all three groups, with five children in each of the TSC and TSO groups and four in the OSC group displaying this process. Nasalization/nasal realization was found for two children in the TSC group, four in the TSO group, and none in the OSC group.

Table 5

The Most Common Phonological/Articulatory Processes Used Consistently (i.e., at Least 20% of Possible Occurrences) in the Three Groups

Process	TSC^* (n = 9)	TSO (n = 9)	OSC (n = 10)
Backing to palatal/velar/uvular	5	5	4
Consonant deletion	3	3	1
Consonant insertion	0	1	0
Cluster reduction	1	1	1
Devoicing	0	1	0
Fricativization	0	1	0
Fronting	2	2	1
Gliding: /l/ to /j/	2	1	1
Gliding: /r/ to /j/	4	2	0
Glottal articulation	1	1	0
H-zation	0	2	0
Lateralization	1	1	1
Lateral /s/	0	1	1
Nasalization/nasal realization	2	4	0
Nasalization of fricatives	0	1	0
Other gliding	1	1	2
Stopping	4	2	1
Voicing	0	1	0

TSC = two-stage repair with closed hard palate; TSO = two-stage repair with open hard palate; OSC = one-stage repair with closed palate.

No significant group differences were found for audible nasal air leakage and hyponasality (Table 6). There was a significant difference among the three groups in hypernasality (Table 6), and post hoc pairwise group comparisons revealed significantly higher degrees of hypernasality for the TSO than for the OSC group (Table 4).

Table 6

Comparisons Among Outcome Measures for Passive Cleft Speech Characteristics at 3 Years of Age (Kruskal-Wallis Test)

	Scale value
	TSC^* (n = 9)^†			TSO^* (n = 9)^†			OSC^* (n = 10)^†
Outcomes	0	1	2	0	1	2	0	1	2	χ²	P
Hypernasality	6	2	1	3	1	5	9	0	1	8.718	.013^‡
Hyponasality	9	0	0	9	0	0	10	0	0	.486	.784
Nasal air leakage	6	3	0	8	1	0	10	0	0	1.549	.461

TSC = two-stage repair with closed hard palate; TSO = two-stage repair with open hard palate; OSC = one-stage repair with closed palate.

†

Number of children.

‡

P < .05.

Discussion

Both primary palatal surgery in one stage and primary palatal surgery in two stages are performed in Sweden. The purpose of the present study was to compare speech and phonology in children treated with three different surgical procedures for palatal closure: a two-stage palatal procedure with hard palate closure at two different time points and a one-stage palatal procedure.

In the present study, early primary palatal surgery in one or two stages did not result in any significant differences in speech production and phonology at age 3 years, but the children with an unoperated cleft in the hard palate at 3 years had significantly poorer speech and phonology. Although the small groups of children studied may have skewed the results, the significant differences between the OSC and TSO groups are presumably stable. The results were also in agreement with the opinion that an unoperated cleft in the hard palate at age 3 years results in poorer expressive speech than if the cleft is closed at an earlier age (Willadsen, 2012). The one-stage procedure in this study seemed to give a better outcome than reported previously (Lohmander, 2011). This may be a result of other variables, such as surgical technique and skill, presence of postoperative fistulas, hearing ability, and general improvement of intervention over time. The latter factor could very well have influenced the results, since the children in the TSC and TSO groups were born between 1997 and 2003, whereas the children in the OSC group were born between 2005 and 2008.

The technique used for palatal closure may be essential for the results. For example, better results were reported at 5 years of age in children born with isolated CP after a two-stage palatal procedure (Persson et al., 2002) than in children of the same age and diagnosis treated with a one-stage closure (Nyberg et al., 2010). However, the techniques for palatal repair differed from those used in the present study. In addition, since the speech of children with isolated CP was evaluated in those studies, direct comparisons with the present study cannot be done. The technique for the two-stage palatal procedure used in the present study was a modification of the two-stage procedure used in Gothenburg. The difference in the procedures is in the nasal layer. In the Gothenburg procedure, the nasal layer is released from the posterior part of the hard palate together with the muscles by an incision and back-cut laterally, with the tensor muscle cut to enable the nasal mucosa with the muscles to rotate posteriorly. The muscles are then sutured in the midline together with the posteriorly based vomer flap. This procedure leaves a defect in the anterior part of the soft palate, between the nasal layer and the hard palate, that is covered only by the oral flaps (Lilja et al., 1996). The release of the nasal layer from the posterior part of the hard palate was not performed in the Scandcleft soft palate repair. Another difference between the procedures was in the muscle dissection for intravelar veloplasty, which was difficult to perform in the Scandcleft project, because the muscles are tiny and poorly defined at 3 to 4 months of age. These modifications used in the Scandcleft soft palate repair were considered to be a disadvantage in comparison to the original procedure for soft palate repair used in Gothenburg.

Surgical caseload is another important factor in success. In a large British study, a higher number of palatal repairs undertaken by the surgeons was associated with a better speech outcome (Williams et al., 1999). In the present study, the children in the OSC group were treated by one surgeon, whereas the TSC and TSO groups were treated by two surgeons. The importance of surgical experience and caseload needs to be further investigated.

There were more children without hearing loss in the OSC group than in the other groups. It is unclear whether the better hearing was a result of fewer instances of otitis media with effusion or better treatment with pressure-equalizing tubes. In the present study, we did not have the full picture of the children's auditory histories, since hearing was assessed at only two to four time points. The possible influence of mild to moderate hearing loss associated with otitis media with effusion on speech and language development in children born with CP cannot be excluded and needs therefore to be further explored (Flynn et al., 2009).

Median percentages for intra- and intertranscriber agreement were in accordance with transcriber agreement in several other studies in the field (e.g., Chapman et al., 2003,2008; Morris and Ozanne, 2003; Scherer et al., 2008). However, the median intertranscriber agreement for two children in the present study was below 60%. This result may have been related to difficulties in understanding the speech in these two children, who had very low PCC-A values (3% and 19.4%, respectively). Furthermore, the interrater agreement for hypernasality was not acceptable, even when a difference of one scale value was accepted, and certainly not after the first two scale values were pooled into one point. The results regarding hypernasality are therefore not reliable. It might have been better to assess hypernasality from connected speech. However, to obtain an equivalent material for all children, we chose to assess single words. In the Scandcleft pilot project, mean values were low for intrarater agreement for hypernasality in high vowels in single words (Lohmander et al., 2009). We assumed that the use of the whole words in a string would increase the validity and also the reliability. Unfortunately, the agreement for hypernasality did not reach an acceptable level. Low intra- and interrater agreement in assessment of hypernasality is a frequently reported problem (e.g., McWilliams et al., 1990; Karling et al., 1993; Keuning et al., 1999; Timmons et al., 2001; Lohmander and Persson, 2008). Even after calibration among listeners, agreement on hypernasality may be low (Karling et al., 1993), but systematic and frequent training may improve agreement (Lee et al., 2009; Sell et al., 2009). Calibration was performed in the current study, but systematic training was not, which might explain the poor agreement on hypernasality.

Backing to palatal/velar/uvular was the most common consistent process (i.e., occurring in at least 20% of all possible occurrences) in all three groups. This was somewhat unexpected, since retracted oral articulation or backing usually has been connected with surgical procedures, including late hard palate repair (e.g., Lohmander and Persson, 2008). The TSO group displayed both more consistent and more sporadic phonological/articulatory processes than the two other groups. This is in accordance with results by Willadsen (2012), who found more cleft speech characteristics and more restricted phonology in children treated with early soft palate closure in combination with a remaining open hard palate at 3 years of age than in their peers treated with early soft palate closure and hard palate closure at 12 months.

Several studies have shown that nonoral and low-pressure consonants are common in babbling in children born with CP and with an unoperated palate at the time of assessment (e.g., Chapman et al., 2001; Scherer et al., 2008). However, children with an early closure of the soft palate, even with an open cleft in the hard palate, display a proportionally high occurrence of oral stops in their babbling (e.g., Willadsen and Albrechtsen, 2006; Lohmander et al., 2011; Willadsen, 2012). Previous findings have shown continuity in development from consonant production in babbling before surgery to consonant production in connected speech at 21 months of age (Chapman et al., 2003). Also, other studies have reported findings on the relationship between consonant production in babbling and early speech development (Chapman et al., 2001; Willadsen and Albrechtsen, 2006; Lohmander and Persson, 2008; Scherer et al., 2008; Lohmander et al., 2011; Willadsen, 2012; Klintö et al., 2013). The occurrence of oral stops at 18 months has been found to be significantly related to the percentage of correct consonants at 3 years of age (Lohmander and Persson, 2008; Klintö et al., 2013). According to these earlier findings, the children in the present study who were treated with early soft palate closure could be assumed to have had a proportionally high occurrence of oral stops in babbling, associated with higher PCC-A at 3 years of age. Similarly, the children treated with a one-stage palatal closure at about 12 months of age could be assumed to have had a proportionally low occurrence of oral stops in babbling, associated with lower PCC-A at 3 years of age. However, this was not the case in this small study. Since babbling was not assessed, we do not know about either the presence of oral stops in babbling or about the continuity with later consonant production in these children. More studies are needed to determine the critical time point for primary palatal closure to create conditions for good development of speech and phonology.

Conclusions

The procedures for early primary palatal surgery in one or two stages used in the present study did not result in any significant differences in speech production and phonology at age 3 years. However, in agreement with recent findings of Willadsen (2012), it could be suggested that a two-stage palatal procedure with delayed closure of the hard palate until the child is 3 years of age will result in speech impairment during the child's early years, at least before hard palate closure.

Footnotes

Acknowledgments

The authors are grateful to speech-language pathologist Liisi Raud-Westberg, who performed the speech analyses.

Explanations of Phonological/Articulatory Processes Used Consistently

Backing to palatal/velar/uvular	An alveolar/dental stop is retracted to velum, for example, /t/ → /k/
Cluster reduction	The number of consonants in a cluster is reduced, for example, /lv/ → /v/
Consonant deletion	A consonant is deleted, for example, /bil/ (car) → /il/
Consonant insertion	An extra consonant is inserted, for example, → /papa/ (daddy) → /pkapa/
Devoicing	A voiced consonant is produced unvoiced, for example, /b/ → /p/
Fricativization	A nonfricative stop is produced with frication, for example, /t/ → /s/
Fronting	A velar stop is produced at an alveolar/dental place, for example, /k/ → /t/
Gliding: /l/ to /j/	The lateral liquid /l/ is produced as /j/ (defined as a glide in Swedish)
Gliding: /r/ to /j/	The vibrant liquid /r/ is produced as /j/ (defined as a glide in Swedish)
Glottal articulation	Oral consonants are retracted to vocal cord level and produced as glottal stops
H-zation	Oral consonants are retracted to vocal cord level and produced as the glottal fricative /h/
Lateralization	A normally nonlateralized consonant as /r/ is produced as the lateral /l/
Lateral /s/	The fricative /s/ is produced with lateral release
Nasalization/nasal realization	Nasal realization of oral consonants, for example, /b/ → /m/
Nasalization of fricatives	An oral fricative as /s/ is produced through the nasal airway
Other gliding	Consonants other than /r/ and /l/ are produced as glides, such as /j/ and /w/
Stopping	A fricative is stopped, for example, /s/ → /t/
Voicing	An unvoiced consonant is produced voiced, for example, /p/ → /b/

References

Aniansson

, Svensson

, Becker

, Ingvarsson

Otitis media and feeding with breast milk of children with cleft palate. Scand J Plast Reconstr Surg Hand Surg. 2002; 36: 9–15.

Chapman

K.L.

, Hardin-Jones

M.A.

, Goldstein

J.A.

, Halter

K.A.

, Havlik

R.J.

, Schulte

Timing of palatal surgery and speech outcome. Cleft Palate Craniofac J. 2008; 45: 297–308.

Chapman

K.L.

, Hardin-Jones

M.A.

, Halter

K.A.

The relationship between early speech and later speech and language performance for children with cleft lip and palate. Clin Linguist Phon. 2003; 17: 173–197.

Chapman

K.L.

, Hardin-Jones

M.A.

, Schulte

, Halter

K.A.

Vocal development of 9-month-old babies with cleft palate. J Speech Lang Hear Res. 2001; 44: 1268–1283.

Farzaneh

, Becker

, Peterson

A.M.

, Svensson

Speech results in adult Swedish patients born with unilateral complete cleft lip and palate. Scand J Plast Reconstr Surg Hand Surg. 2008; 42: 7–13.

Flynn

, Moller

, Jonsson

, Lohmander

The high prevalence of otitis media with effusion in children with cleft lip and palate as compared to children without clefts. Int J Pediatr Otorhinolaryngol. 2009; 73: 1441–1446.

Friede

Maxillary growth controversies after two-stage palatal repair with delayed hard palate closure in unilateral cleft lip and palate patients: perspectives from literature and personal experience. Cleft Palate Craniofac J. 2007; 44: 129–136.

Friede

Two-stage palate repair. In: Losee

, and Kirschner

R.E.

, eds. Comprehensive Cleft Care. New York: McGraw Hill; 2009: 413–429.

Friede

, Lilja

, Lohmander

Two-stage palatal surgery with early veloplasty and delayed hard palate repair: a balanced view on speech and midfacial growth outcome. In: Berkowitz

, ed. Cleft Lip and Palate. Berlin: Springer; 2013: 432–433.

10.

International Phonetic Association. Handbook of the International Phonetic Association. Cambridge, UK: Cambridge University Press; 1999.

11.

Karling

, Larson

, Leanderson

, Henningsson

Speech in unilateral and bilateral cleft palate patients from Stockholm. Cleft Palate Craniofac J. 1993; 19: 261–267.

12.

Keuning

K.H.

, Wieneke

G.H.

, Dejonckere

P.H.

The intrajudge reliability of the perceptual rating of cleft palate speech before and after pharyngeal flap surgery: the effect of judges and speech samples. Cleft Palate Craniofac J. 1999; 36: 328–333.

13.

Klintö

, Salameh

E.K.

, Flynn

, Olsson

, Svensson

, Lohmander

Phonology in Swedish-Speaking 3-Year-Olds Born with Cleft Lip and Palate and the Relationship with Consonant Production at 18 Months. Accepted for publication in International Journal of Language and Communication Disorders 2013.

14.

Klintö

, Salameh

E.K.

, Svensson

, Lohmander

The impact of speech material on speech judgement in children with and without cleft palate. Int J Lang Commun Disord. 2011; 46: 348–360.

15.

Lee

, Whitehill

T.L.

, Ciocca

Effect of listener training on perceptual judgement of hypernasality. Clin Linguist Phon. 2009; 23: 319–334.

16.

Lilja

, Friede

, Johanson

Changing philosophy of surgery of the cleft lip and palate in Göteborg, Sweden. In: Berkowitz

, ed. Cleft Lip and Palate. Vol 2. San Diego: Singular; 1996: 155–170.

17.

Lohmander

Surgical intervention and speech outcome in cleft lip and palate. In: Howard

, Lohmander

, eds. Cleft Palate Speech: Assessment and Intervention. Chichester: Wiley-Blackwell; 2011: 55–85.

18.

Lohmander

, Friede

, Elander

, Persson

, Lilja

Speech development in patients with unilateral cleft lip and palate treated with different delays in closure of the hard palate after early velar repair: a longitudinal perspective. Scand J Plast Reconstr Surg Hand Surg. 2006; 40: 267–274.

19.

Lohmander

, Friede

, Lilja

Long-term, longitudinal follow-up of individuals with unilateral cleft lip and palate after the Gothenburg primary early veloplasty and delayed hard palate closure protocol: speech outcome. Cleft Palate Craniofac J. 2012; 49: 657–671.

20.

Lohmander

, Olsson

Methodology for perceptual assessment of speech in patients with cleft palate: a critical review of literature. Cleft Palate Craniofac J. 2004; 41: 64–70.

21.

Lohmander

, Olsson

, Flynn

Early consonant production in Swedish infants with and without unilateral cleft lip and palate and two-stage palatal repair. Cleft Palate Craniofac J. 2011; 48: 271–285.

22.

Lohmander

, Persson

A longitudinal study of speech production in Swedish children with unilateral cleft lip and palate and two-stage palatal repair. Cleft Palate Craniofac J. 2008; 45: 32–41.

23.

Lohmander

, Willadsen

, Persson

, Henningsson

, Bowden

, Hutters

Methodology for speech assessment in the Scandcleft project—an international randomized clinical trial on palatal surgery: experiences from a pilot study. Cleft Palate Craniofac J. 2009; 46: 347–362.

24.

McReynolds

L.V.

, Elbert

Criteria for phonological process analysis. J Speech Hear Disord. 1981; 46: 197–204.

25.

McWilliams

B.J.

, Morris

H.L.

, Shelton

R.L.

Diagnosis of disorders of phonation and resonance. In: McWilliams

B.J.

, Morris

H.L.

, Shelton

R.S.

, eds. Cleft Palate Speech. Philadelphia: BC Decker; 1990: 311–319.

26.

Morris

, Ozanne

Phonetic, phonological, and language skills of children with a cleft palate. Cleft Palate Craniofac J. 2003; 40: 460–469.

27.

Nyberg

, Raud-Westberg

, Neovius

, Larson

, Henningsson

Speech results after one-stage palatoplasty with or without muscle reconstruction for isolated cleft palate. Cleft Palate Craniofac J. 2010; 47: 92–103.

28.

Persson

, Elander

, Lohmander-Agerskov

, Söderpalm

Speech outcomes in isolated cleft palate: impact of cleft extent and additional malformations. Cleft Palate Craniofac J. 2002; 39: 397–408.

29.

Persson

, Willadsen

, Havstam

, Davies

, Kisling-Mller

, Kjll

, Lundberg

, Pedersen

, Schps

, Lohmander

The Scandcleft randomized trials: hypernasality in 5-year-olds with UCPL. Presented at the 12th International Congress on Cleft Lip/Palate and Related Craniofacial Anomalies; May 2013; Orlando, Florida, USA.

30.

Peterson-Falzone

S.J.

The relationship between timing of cleft palate surgery and speech outcome: what have we learned, and where do we stand in the 1990s?

Semin Orthod. 1996; 2: 185–191.

31.

Pigott

R.W.

, Albery

E.H.

, Hathorn

I.S.

, Atack

N.E.

, Williams

, Harland

, Orlando

, Falder

, Coghlan

A comparison of three methods of repairing the hard palate. Cleft Palate Craniofac J. 2002; 39: 383–391.

32.

Roberts

C.T.

, Semb

, Shaw

W.C.

Strategies for the advancement of surgical methods in cleft lip and palate. Cleft Palate Craniofac J. 1991; 28: 141–149.

33.

Roberts

, Hunter

, Gravel

, Rosenfeld

, Berman

, Haggard

, Wallace

Otitis media, hearing loss, and language learning: controversies and current research. J Develop Behav Pediatr. 2004; 25: 110–122.

34.

Scherer

N.J.

, Williams

A.L.

, Proctor-Williams

Early and later vocalization skills in children with and without cleft palate. Int J Pediatr Otorhinolaryngol. 2008; 72: 827–840.

35.

Sell

Issues in perceptual speech analysis in cleft palate and related disorders: a review. Int J Lang Commun Disord. 2005; 40: 103–121.

36.

Sell

, John

, Harding-Bell

, Sweeney

, Hegarty

, Freeman

Cleft audit protocol for speech (CAPS-A): a comprehensive training package for speech analysis. Int J Lang Commun Disord. 2009; 44: 529–548.

37.

Semb

Scandcleft randomized trials of primary surgery for unilateral cleft lip and palate. Presented at WHO meeting on International Collaborative Research on Craniofacial Anomalies; November 2001; World Health Geneva, Switzerland.

38.

Sommerlad

B.C.

Surgery of the cleft palate: repair using the operating microscope with radical muscle retropositioning—the GostA approach. B-ENT. 2006; 2(Suppl 4): 32–34.

39.

Spauwen

P.H.

, Goorhuis-Brouwer

S.M.

, Schutte

H.K.

Cleft repair: Furlow versus von Langenbeck. J Craniomaxillofac Surg. 1992; 20: 18–20.

40.

Timmons

M.J.

, Wyatt

R.A.

, Murphy

Speech after repair of isolated cleft palate and cleft lip and palate. Br J Plast Surg. 2001; 54: 377–384.

41.

Van Demark

D.R.

, Gnoinski

, Hotz

M.M.

, Perko

, Nussbaumer

Speech results after the Zürich approach in the treatment of unilateral cleft lip and palate. Plast Reconstr Surg. 1989; 83: 605–613.

42.

Van Lierde

K.M.

, Monstrey

, Bonte

, Van Cauwenberge

, Vinck

The long-term speech outcome in Flemish young adults after two different types of palatoplasty. Int J Pediatr Otorhinolaryngol. 2004; 68: 865–875.

43.

Willadsen

Influence of timing of hard palate repair in a two-stage procedure on early language development in Danish children with cleft palate. Cleft Palate Craniofac J. 2012; 49: 574–595.

44.

Willadsen

, Albrechtsen

Phonetic description of babbling in Danish toddlers born with and without unilateral cleft lip and palate. Cleft Palate Craniofac J. 2006; 43: 189–200.

45.

Williams

A.C.

, Sandy

J.R.

, Thomas

, Sell

, Sterne

J.A.

Influence of surgeon's experience on speech outcome in cleft lip and palate. Lancet. 1999; 13;354: 1697–1698.

46.

Zarcone

J.R.

, Rodgers

T.A.

, Iwata

B.A.

, Rourke

D.A.

, Dorsey

M.F.

Reliability analysis of the motivation assessment scale: a failure to replicate. Res Dev Disabil. 1991; 12: 349–360.