The clinical and cost effectiveness of cochlear implants (CIs) as a treatment for children with severe to profound hearing loss is universally accepted (1,2). Forli et al.’ systematic review considered over 929 international research articles to examine effectiveness and guidelines/policies on cochlear implant procedures (3). Three themes emerged; bilateral versus unilateral CI use, the application of CIs for children with additional diagnoses/disabilities, and of specific relevance to the current study, the association between optimum communication outcomes and earlier access to cochlear implantation, that is, the “earlier is better” argument. They summarized 22 research studies, of which 7 specifically detailed outcomes for children implanted younger than 12 months. Problematic to their review, and to this field in general, was the range of methodologies employed, and the variability in test materials. More recent findings, since Forli's review, in speech perception, language development, and speech production for children implanted younger than 12 months are described here.
In 1990, the USA's Food and Drug Administration's (FDA) guidelines approved cochlear implantation in children aged greater than 24 months who have bilateral profound sensorineural hearing loss (>90 dB HL). These criteria were lowered to 18 months in 1998, with a further lowering to 12 months in 2000. Surgery in infants younger than 12 months, in countries bound by FDA guidelines, constitutes “off-label” use of CIs by the Food and Drug Administration. Data that support parents’ decision to proceed with CI surgery for their infants is therefore derived from countries other than North America, and some specific research studies within the United States. For the group aged 6 months to 12 months, literature arguing that “the benefits of early implantation in carefully selected infants may mitigate the concerns” (p. 376) has been available since 2004 (4). CI surgery in children younger than 12 months can be completed with minimal or no increased surgical risk with long-term follow-up reported for up to 10 years (5–11). Furthermore, in Europe, Australia, and South Africa (countries not bound by the FDA guidelines) infants as young as 4 months have received cochlear implant(s). Parents tend to go ahead with surgery under 6 months in two scenarios; first, when siblings have already received CI(s) (i.e., parents are well aware of hearing tests, procedures, and benefits of CIs), and second, for medical reasons such as MRI evidence of precursors to cochlear ossification postmeningitis.
Due to the difficulties in testing speech understanding for young children, early published research regarding potential benefits of CIs younger than 12 months often included checklists and categories of performance, or researchers waited until open-set word testing was possible. Colletti et al. (5) used categories of auditory performance (CAP) to show significant advantages for 10 children who received CIs younger than 12 months. Using a modified speech perception checklist (IT-MAIS) in interviews with parents, Robbins et al. (12) reported data from early clinical trials in North America. Forty-five children who received implants between 12 and 17 months demonstrated faster rates of growth than children who received implants at older ages (n = 32 implanted between 19 and 23 mo, and n = 30 children implanted between 24 and 36 mo). Their data, however, included no children who received CIs younger than 12 months. Waltzman and Roland (11) reported that cochlear implant surgery was feasible and IT-MAIS results for 18 children improved with greater device experience; however, they only had open-set word understanding results data for five patients.
Valencia et al. (8) also used the IT-MAIS to demonstrate speech understanding in 15 infants who received CIs under 12 months. The authors acknowledged that they did not set out to prove long-term benefits of CIs and, therefore did not have a comparison group of children who received CIs older than 12 months. Holt and Svirsky (13) and Tajudeen et al. (14) used rate of word understanding measures to examine the influence of age at implant. Holt and Svirsky (13) used a “follow-the-instructions-task” during play with a Mr Potato Head (as a modified open-set word understanding test) and did not find that age at implantation significantly influenced rate of word recognition in 96 children. Their study, however, included only six children who received implants younger than 12 months. Tajudeen et al. (14) demonstrated higher speech understanding in 35 children who received implants younger than 12 months compared with older groups. Although most of their data were obtained from the child participants’ completion of the LNT-easy word lists, the authors acknowledged that, at some time points, predictive functions were based on linear regression analyses from other measures including GASP words, MLNT hard words, common phrases sentences in quiet, LNT hard words, and PBK words scores.
Further examination of the speech perception scores (open-set, live voice, audition alone) achieved by children using cochlear implants is still warranted. In a prospective longitudinal study, children with bilateral CIs were age matched with children with normal hearing/typical development. Both groups completed the LittlEARS parent questionnaire and single-syllable word tests (administered live voice) at regular intervals postimplant (15). Within 9 months postimplant, child participants using CIs matched their hearing peer's LittlEARS performance; the youngest at CI surgery demonstrated the highest scores in the CI group. The mean single-syllable word score for CI users was 80.5% (SD 12.7) where the hearing controls scored 91.9% (SD 6.5). Leigh et al. (16) demonstrated a mean open-set (PBK and CNC) word score of 54% and phoneme score of 78% in 80 unselected children who received cochlear implants younger than 3 years. These scores exceeded mean scores achieved by children using hearing aids with pure-tone averages (PTA) in the severe and profound range and were not statistically different from scores achieved by children using hearing aids with PTAs in the moderate range.
Before concluding this review of speech perception data, it is worth considering studies that did not find a relationship between age at implant and speech perception outcomes. Wie et al. (17) described outcomes for all 79 prelingually deaf children implanted in Norway. The average age at implantation was 50 months and only one child was implanted before 2 years of age. Progress on their compiled questionnaire (derived from the meaningful auditory integration scale (MAIS), meaningful use of speech scale (MUSS) (18), categories of auditory performance (CAP), and listening progress (LIP) (19)) was not predicted by age at CI. There was also no relationship found between age at implant and speech perception, language or speech production outcomes in a large study of 181 North American and Canadian children (20). The child participants had a mean age at implant of 3 year 5 months (range, 1;8–5;4; SD, 10 mo). It could be argued that the child participants in both studies (17,20) did not receive CI(s) at an age that was young enough to influence a significant difference in speech perception outcomes. Leigh et al. (21) directly compared open-set word understanding in 27 children who received CIs younger than 12 months (word score 51%, phoneme score 79%), and scores were not significantly different to 68 children who received CIs between 13 and 24 months (word score 52%, phonemes 78%). Black et al. (22) found no evidence that age at implant was associated with optimum speech perception. They collected data (n = 133) using the Categories of Auditory Performance Index (CAPI). Their data, however, were confounded by the inclusion of children with progressive course of hearing loss. These children typically develop good perception and language skills before the hearing deterioration and tend to receive CIs late, after hearing has deteriorated to profound levels.
In summary, the evidence supporting critical periods for speech perception development using data from children receiving CIs is unclear. It may be possible for children to develop speech perception skills equally well if they have access to CIs before or after 12 months of age. Data from hearing infants, however, show an important link between perception and language. Tsao et al. (23) demonstrated that speech perception at 6 months of age (in 28 hearing infants) was correlated with language at 2 years of age. The present study argues that access to speech perception via cochlear implants before 12 months may be a necessary prerequisite to spoken language and intelligible speech.
Numerous studies have demonstrated that mean language scores for children who received CIs younger than 2.5 years were higher than mean scores achieved for children who received CIs older than 2.5 years (24–28). What remains to be proven is whether significant advantages exist for children who proceed with CIs younger than 12 months. As is the case for testing speech perception, the careful selection of test materials that are age and stage appropriate is important when demonstrating language outcomes.
Connor et al. (29) demonstrated that age at implant predicted rates of vocabulary growth (PPVT) in 100 children aged 12 months to 10 years at CI. Tomblin et al. (30) also demonstrated more rapid growth in language for children implanted younger. They examined MCDI (a checklist of 300 questions completed by the parent) and PLS-3 (standardized test of receptive and expressive language) growth rates pre- and postimplant in 29 children with a mean age at CI of 1 year and 9 months (range, 10.5 mo–3 yr 4 mo; SD, 7 mo). Not all children completed all tests at all intervals, and only two child participants were actually aged under 12 months at CI switch on. Despite these limitations, the authors concluded “that earlier access to sound alters some fundamental systems involved in language learning in such a way that the long-term trajectory for development is altered” (Tomblin et al. (30), p. 864). They proposed a model whereby the child experiences a shorter interval of “total” deafness during which he or she is falling behind, and a steeper gradient of language growth (than that demonstrated by their hearing peers) postcochlear implant to “catch up” or “close the gap.” Both components, the short auditory deprivation period and the accelerated rate of language acquisition, are needed for the child to demonstrate age appropriate language by school entry.
The concept of cumulative practice is also relevant here (31). The child's learning and practice within meaningful interactions begets further learning and practice. Starting this positive cycle early is consistent with the notion of working within the design specifications of language (32). Nicholas and Geers (33) summarized that the hearing infant's inherent design includes;
preference for human speech over rhesus vocalizations by 3 months of age (Vouloumanos, Hauser, Werker, & Martin, 2010), the development of word segmentation abilities between 7.5 and 10.5 months of age (Jusczyk, 2002), the ability to associate words with salient persons, common objects, and body-parts at 6 months (Bergelson & Swingley, 2012; Tincoff & Jusczyk, 1999; Tincoff & Jusczyk, 2012), and recognition of change in the identity of a speaker at 7 months (Johnson, Westrek, Nazzi, & Cutler, 2011) (cited in Nicholas & Geers, 2013, p. 533).
Recent literature has shown that children implanted younger than 12 months tend to have better language outcomes than children implanted later. These data may be grouped into the following domains: improved language trajectories (6,7,34,35), improved word learning (36,37), standard score equivalents approaching normal hearing peers (15,38), less language delay (21), and language scores that are equivalent to chronological age (6).
In two separate studies, Ching et al. (34) and Dettman et al. (7) showed statistically better rates of receptive and expressive development for Australian children implanted younger than 12 months (Ching et al., n = 10; Dettman et al., n = 11) compared with children who received CIs later (Ching et al., n = 21, Dettman et al., n = 36). Children who received CIs younger than 12 months of age in both studies demonstrated language growth rates equivalent to their hearing peers.
Cuda et al. (36) demonstrated greater vocabulary development for children implanted younger using the MacArthur-Bates Communicative Development Inventory (MCDI). Checklists were completed by parents when children were aged 36 months. Of their 30 child participants (mean age at CI 11.8 mo; range, 8–17 mo; SD, 3.2 mo), 14 children were younger than 12 months at CI. Better total expressive vocabulary, length of utterance, and sentence complexity on the MCDI were associated with younger age at CI. Direct comparisons, however, were limited by the differing length of device experience in these groups. That is, a child implanted at 17 months has 19 months to develop language, whereas a child implanted at 8 months has 28 months to develop language, if the MCDI is completed at 36 months.
In Wie (15) the majority of children demonstrated expressive and receptive language skills within the normative range after 12 to 48 months; this proportion increased with implant experience. Early age at CI was associated with better language scores on the Mullen Scale of Early Learning, and the Minnesota Child Development Index during the early postimplant period (particularly for expressive language), but the advantages lessened over time. According to Tomlin's proposed language model, data supporting CIs in children younger than 12 months would need to demonstrate two facts; that there is less time for gaps to develop, and that exceptional rates of development can be maintained postimplant to facilitate language acquisition. Leigh et al. (21) demonstrated that the rate of receptive language growth for 32 children who received CIs under 12 months (0.90, where 1.0 is normal) was not significantly different to 42 children who received CI between 13 and 24 months (0.92), but the degree of delay (roughly equivalent to the age at implant) was significantly different. When these children were followed for 3 years and completed the PPVT, there were significant differences in their standard scores. The mean receptive vocabulary standard score (SS) for 21 children who received CIs under 12 months (SS 90.8) was within the normal range for hearing peers (mean, 100; ±SD, 15; range, 85–115) and statistically, was significantly better than the standard score achieved for 40 children who received CIs between 13 and 24 months (SS 80.6).
Both May-Mederake (38) and Colletti et al. (6) demonstrated that a greater proportion of children who received CIs younger than 12 months demonstrated language performance within the same range as their hearing peers, compared with cohorts who received CIs when older than 12 months. May-Mederake investigated 11 children who received CIs under 12 months, compared with n = 5 who were aged between 12 and 18 months at CI, and n = 5 aged over 18 months at CI, but not all children completed all test measures. Colletti et al. (1,6) reported long-term data for children completing the PPVT. For the n = 11 who were younger than 12 months (2–11 mo) at CI, the language age equivalent at a chronological age of 10 years was 9.5 years (SD ± 0.3) which was within the normative range. In contrast, the mean age equivalent for n = 13 who received CIs aged 12 to 23 months (and aged 10 yr at test) was 8.3 years (SD ± 0.5). For n = 19 who were aged 24 to 35 months at CI (and 10 yr at test), it was 7.8 years (SD ± −1.0) and for n = 25 who were aged 72 to 83 months (and 10 yr at test), the language equivalent was 5.8 years (SD ± 1.2). (It is not known why there were no children aged between 35 and 72 mo in their study.) The authors concluded that these significant receptive vocabulary delays in the older at implant groups represented a substantial cost burden in terms of long-term educational support.
Black et al. (22) demonstrated marginal evidence that age at implant was associated with optimum language outcomes (measured by a range of language tests including the PLS-4, CELF-3, PPVT-3, and EVT) but, as mentioned previously, their inclusion of children with acquired and progressive hearing loss losses confounded this enquiry.
To conclude this review of limitations with current language outcome data, standardized prospective protocols are lacking. In order to argue the case for infants to have earlier access to hearing and to shift current FDA guidelines, large prospective multi-centre clinical studies which employ consistent measurement tools to examine long term communication outcomes will be required. While waiting for such clinical trials to occur, however, significant numbers of children with hearing loss may miss their opportunities for optimum speech perception development, with cascading consequences to their spoken language development (39).
Speech production may be described in terms of onset of vocal development, quantity (volubility), and quality (e.g., percentage accuracy of vowels, consonants, and clusters). Speech production data may be derived by direct observation/sampling and/or by completion of standardized tests of articulation. Standardized tests may be reported in terms of phoneme accuracy, standard scores, or percentile ranks. Research has shown that progress in vocal development may be comparable to hearing peers if children receive CIs before 3 years of age and are enrolled in oral intervention programs (40,41). Children using CIs actually demonstrated more rapid development in speech after activation than their hearing peers, suggesting optimum “readiness” for speech development. Onset of babbling 4 months after activation of the CI (42), more rapid development than hearing peers (40), and babbling development not significantly different to hearing peers (5,43) have been reported for children who received CIs younger than 12 months. The importance of early vocal developments is its concurrent role in phonological and expressive vocabulary development (44–46). Connor et al. (29) reported better vowel and consonant accuracy for children who had earlier CIs, but did not have any children younger than 12 months at implant. A major U.S study involving 181 children who received CIs under 6 years did not find age at implant to predict speech production, but as mentioned earlier with regard to speech perception and language, the data did not include any children with CIs under 12 months (47).
Overall, there are less data available which directly compare speech intelligibility outcomes for children receiving cochlear implants younger than 12 months with older cohorts. This may be due to the wait required before children can demonstrate speech production skills on standardized tests of articulation. For example, Colletti et al. (6) used a rating of speech intelligibility, rather than articulation test results. Black et al. (22) reported that their retrospective study collected articulation test data (GFT-2 and DEAP) on just n = 17, from the potential group of 174 medical records. The age at implant data was not specified and the authors did not discuss the speech production results further. Leigh et al. (21) waited until child participants were a least 2 years postimplant to complete the Diagnostic Evaluation of Articulation and Phonology (DEAP). The 16 children who received CIs younger than 12 months achieved significantly higher speech production scores than 16 children who received CIs at 13 to 24 months. Both groups displayed speech that was, however, significantly poorer than their hearing peers.
A summary of the current literature examining communication outcomes for children who receive CIs younger than 12 months is in the Appendix. The present study aimed to describe the potential range of performance for children receiving cochlear implants on standardized measures of communication. Data from Australian centers are uniquely placed to describe long-term outcomes as these children were the first to receive CIs at under 12 months of age, and results from a battery of tests have been collected via prospective clinical assessment protocols for up to 15 years postimplant. As there were significant numbers of children contributing to these data, this study aimed to examine mean performance for cohorts implanted between 6 and 12 months, 13 and 18 months, 19 and 24 months, 24 months and 42 years, and 43 and 60 months of age. It was hypothesized that there would be a relationship between age at implant and outcome measure for speech perception, language, and speech production. Specifically, it was hypothesized that the group who received cochlear implants younger than 12 months would be more likely to demonstrate language skills within the normative range for their hearing peers.
Three Australian early intervention/cochlear implant centers were selected for this study as they used a prospective assessment battery that included speech perception, language, and speech production measures. They completed assessments with the children at school entry and late primary/early secondary school time points. Data were entered into their respective databases at the time of assessments. After ethical approval, data from three databases were pooled to examine outcomes for over 800 children.
Children were included in this study if they had a congenital stable severe to profound bilateral hearing loss, demonstrated cognitive skills within the normal and borderline normal range, and had received cochlear implant(s) younger than 6 years of age. No further exclusion criteria were used; therefore, this diverse group implanted between 1990 and November 2014 used a range of communication approaches (oral–aural, auditory verbal, and sign language approaches) and had a range of aetiologies and comorbid conditions. The children used the Nucleus multichannel device (Cochlear Ltd., Sydney, Australia) and speech processors that were current at the time of testing. All children were upgraded to new speech processors as soon as these became available (via government funding). Children attended for regular mapping and communication review appointments at their respective centers. There were 403 children who met these criteria; Group 1 comprised 151 children aged younger than 12 months at CI (33.8%). Group 2 to 5 age cohorts are outlined in Table 1.
The age divisions were chosen to determine whether significant differences in communication outcomes could be observed in groups implanted younger than 12 months (Group 1), younger than 18 months (Group 2), and younger than 24 months (Group 3). This study had sufficient numbers in each group to afford statistical power. In contrast, Connor et al. (29) could not examine outcomes for groupings between 12 and 21 months as they had only four children in that group. Other studies also lacked sufficient numbers in the youngest groups to enable statistical comparison, used categories of performance to describe outcomes (6–8,10,11,13,15,21,34,36), or included children with progressive course of hearing loss that confounded their findings (22). The next division, from 25 months to 42 months (3.5 yr) (Group 4), was chosen due to Sharma et al.'s (48) findings regarding differences in cortical responses for children who received CIs younger than 3.5 years compared with brain responses obtained for children older than 3.5 years at CI. The final division was from 43 to 50 months (5.99 yr) (Group 5). The decision to include children with an age of implant up to 6 years was in response to the diversity of factors affecting children who proceed to implantation after 6 years of age. In Australia, this late-to-implant group often includes refugees from other countries, children with multiple medical issues, children/adolescents who have used Auslan communication with no auditory stimulation, and children with significant residual hearing in one or both ears. To control for these and other variables, children with progressive hearing loss acquired hearing loss and children receiving CIs after age 6 years were excluded. The above group divisions afforded balanced numbers in the five age groups for speech perception and language tests; a different grouping was used for the speech production data.
Using results from a range of educational psychology tests (49–56) relevant to the age and developmental stage of the child, two clinicians who were familiar with the child rated the child's cognitive skills. (See Dettman et al. (57) for further information on use of this rating.) The children's cognitive skills were rated as 0 = accelerated (n = 3), 1 = within the normal range (n = 280), and 2 = borderline, low average (n = 42). Children who obtained a rating of 3 = mild, 4 = moderate, and 5 = severe cognitive delay/impairment were excluded from this study. There were 78 children (19.3%) for whom no cognitive rating was available; their data were still included in this analysis. Despite the exclusion of children with significant cognitive delay, there were still some differences across the groups for cognitive ratings as demonstrated by one-way ANOVA. In particular, mean ratings of cognitive skills for children who received CIs younger than 12 months were higher than Groups 2, 4, and 5 (Table 2).
This difference in cognitive skills, which could confound examination of the influence of age at implant on outcomes, was accounted for by using multiple regression analysis to assess the outcome variables.
All assessments were completed by a certified speech pathologist or audiologist following the test guidelines in a quiet room/audio booth at a level of approximately 65 dB SPL, seated on the best side for listening, using the child's everyday best listening condition. This included cochlear implant alone, bimodal (cochlear implant and hearing aid), and bilateral cochlear implant conditions.
Open-set monosyllabic word (OSW) recognition was tested using CNC words (58). Scores for correct phonemes (out of 150) (OSWph) and correct words (out of 50) (OSWw) were expressed as a percentage. There were 125 children who completed testing at school entry and 81 children who completed testing at late primary/early secondary school time points. Open-set sentence recognition was tested using BKB sentences (59). Scores for correct key words (out of 50) (OSS) were expressed as a percentage. There were 123 children who completed testing at school entry and 89 children who completed testing at late primary/early high school time points. The number of children, mean age at test, mean device experience, mean word (OSWw) and phoneme (OSWph) scores, and mean sentence scores (OSS) for Groups 1 to 5 are in Table 3.
All assessments were completed by a certified speech-language pathologist following the manual guidelines for basal and ceiling scoring, in a quiet room seated in the best position for the child.
PLS-4 and 5
Receptive and expressive language skills were assessed using the Preschool Language Scale-Fourth and Fifth Editions (PLS-4 and PLS-5) (60,61) that are suitable for use from birth to 6 years 11 months. Expressive, receptive, and total language standard scores were obtained for 95 children at school entry; mean age at test 5.4 years (range, 4.0–6.8; SD, 0.5), mean device experience 3.8 years (range, 0.9–5.7; SD, 1.0). The mean PLS standard score achieved by the group was 78.8 (range, 50 to 135; SD, 23.5). The number of children, mean age at test, mean device duration, mean PLS standard score, number and percentage within the normal range for Groups 1 to 5 are in Table 4.
Receptive vocabulary skills were assessed using the Peabody Picture Vocabulary Test (PPVT) that is suitable for use from 2 years to 90 years. The current version of the PPVT was used as soon as it became available; data collected since 1990 include PPVT-3 and PPVT-4 versions (62,63). At school entry 207 children completed the PPVT. The mean age at test was 5.6 years (range, 3.7–8.6; SD, 0.87) with a mean postimplant duration of 3.4 years (range, 0.5–5.77; SD, 1.1 yr). The mean standard score achieved by the group was 75.5 (range, 30–138; SD, 26.1). The number of children, mean age at test, device experience, PPVT standard score, number, and percentage within the normal range for Groups 1 to 5 are in Table 5.
Receptive and expressive language skills were assessed for 122 children using the Clinical Evaluation of Language Fundamentals (CELF). Data from the CELF Preschool – Second Edition, Australian and New Zealand Standardised Edition (CELF P-2 Australian and New Zealand) (64), suitable for children aged 3 to 6 years 11 months, and the CELF – Fourth Edition, Australian Standardised Edition (CELF-4 Australian) (65), suitable for ages 5 to 21 years, were compiled. The CELF provides receptive, expressive, and total language standard scores. The total language standard scores for 122 children were analyzed. The mean age at test was 8.02 years (range, 3.08–15.3; SD, 2.22) with a mean postimplant duration of 6.03 years (range, 1.26–10.8; SD, 2.02 yr). The mean standard score achieved by the group was 77.8 (range, 40–140). The number of children, mean age at test, mean device experience, CELF total standard scores, and proportion within the normal range, for Group 1 to 5 are in Table 6.
The Diagnostic Evaluation of Articulation and Phonology (DEAP) (66) was completed with 76 children at school entry. The child's responses to picture stimuli from the standardized articulation test were transcribed and analyzed using Computer Assisted Speech and Language Analysis software (CASALA) (67). This generated a percentage accuracy for vowels, consonants, clusters, phoneme accuracy, and total word intelligibility (Table 7).
Regression analysis was used to examine whether age-at-implant and/or cognitive skills accounted for variance in speech perception, language, and speech production outcome measures. ANOVA and Tukey pairwise comparison was used to determine whether particular age groups demonstrated significant differences in speech perception scores, language standard scores, and/or vowel, consonant, cluster, and word accuracy over other age groups. Percentage scores for the speech perception data were arcsine transformed before analysis to equalize variance across the range. Finally, cross tabulation and χ 2 tests were used to examine whether a greater percentage of children from Groups 1 to 5 demonstrated language performance within the normative range by school entry. For the speech production data, there were smaller numbers of children who completed the DEAP, and no data for children aged over 3.5 years at implant (Group 5), so data were collapsed into two groups for analysis. Two sample t tests were used to examine whether Group 1 demonstrated greater speech production accuracy than older children. A significance level of p < 0.05 was used in all statistical analyses.
Speech Perception: Results at School Entry
Regression analysis showed a significant relationship between age at implant and open-set word (F = 10.67, p < 0.001), phoneme (F = 9.49, p < 0.003), and sentence (F = 11.64, p < 0.001) scores achieved at school entry. The percentage of the variance in word, phoneme, and sentence scores accounted for by age at implant was 9%, 7%, and 9%, respectively. Regression analysis also showed a significant relationship between cognitive skills and open-set word (F = 9.78, p < 0.002), phoneme (F = 19.42, p < 0.001), and sentence (F = 16.11, p < 0.001) scores achieved at school entry. The percentage of the variance in word, phoneme, and sentence scores accounted for by cognitive skills was 8%, 15%, and 12%, respectively. These two factors accounted for 17 to 20% of the variance in open-set speech perception at school entry. ANOVA with Tukey pairwise comparison for arcsine transformed data showed that the mean word, phoneme, and sentence scores from Groups 1, 2, and 3 were significantly better than Groups 4 and 5 (Fig. 1).
Speech Perception: Results at Late Primary/Early Secondary School
Regression analysis showed a significant relationship between age at implant and open-set word (F = 38.54, p < 0.000), phoneme (F = 31.20, p < 0.000), and sentence (F = 11.29, p < 0.001) scores achieved at late primary/early high school testing. The percentage of the variance in word, phoneme, and sentence scores accounted for by age at implant was 33%, 30%, and 13%, respectively. Regression analysis also showed a significant relationship between cognitive skills and open-set word (F = 4.43, p < 0.039), and sentence (F = 7.50, p < 0.008) scores achieved at school entry. (The relationship was not significant for phoneme scores.) The percentage of the variance in word and sentence scores accounted for by cognitive skills was 5% and 7%, respectively. These two factors, age at implant and cognitive skills, accounted for 20 to 38% of the variance in open-set speech perception at late primary/early secondary school testing. ANOVA with Tukey pairwise comparison for arcsine transformed data showed that the mean word, phoneme, and sentence scores from Groups 1, 2, and 3 were significantly better than Groups 4 and 5 (Fig. 2).
Language: PLS-4 and 5
Regression analysis showed a significant relationship between age at implant (F = 13.39, p < 0.001) and cognitive skills (F = 12.95, p < 0.001) and PLS standard scores achieved at school entry. The percentage of the variance in PLS scores accounted for by age at implant was 18% and by cognitive skills 16%. These two factors accounted for 34% of the variance in PLS standard scores at school entry. ANOVA with Tukey pairwise comparison showed that the mean PLS standard score from Group 1 was significantly greater than Groups 2, 4, and 5. The Group 3 mean was also significantly greater than Group 4 (Fig. 3). Cross tabulation and χ 2 analysis supported the hypothesis that a greater percentage of children in Group 1 (64%) obtained PPVT standard scores within the normal range at school entry (χ 2 = 26.509, df = 4, p < 0.001) than Groups 2, 3, 4, and 5.
Regression analysis showed a significant relationship between age at implant (F = 69.99, p < 0.001) and cognitive skills (F = 11.16, p < 0.001) and PPVT standard scores achieved at school entry. The percentage of the variance in PPVT scores accounted for by age at implant was 27%, and by cognitive skills 5%. These two factors accounted for 32% of the variance in PPVT standard scores at school entry. ANOVA with Tukey pairwise comparison showed that the mean PPVT standard score from Group 1 was significantly better than Group 2. Both Groups 1 and 2 were significantly better than Groups 3, 4, and 5 (Fig. 4). Cross tabulation and χ 2 analysis supported the hypothesis that a greater percentage of children in Group 1 (80%) obtained PPVT standard scores within the normal range at school entry (χ 2 = 73.399, df = 4, p < 0.001) than Groups 2, 3, 4, and 5.
The potential impact of the two significant factors, age at implant, and cognitive skills, is best demonstrated in the range of results on the PPVT at school entry (Fig. 5). For children with normal cognitive function (circles) the data show more children within the normal range if they received CIs under 18 months, with the greatest concentration within the normal range if they received CIs under 12 months. For children with an additional diagnosis of borderline cognitive delay/impairment or low average performance on educational psychology testing (diamonds), only two demonstrate receptive vocabulary within the normal range. The dashed regression line indicates less impact of age at implant for this group.
Regression analysis showed a significant relationship between age at implant (F = 22.66, p < 0.001) and cognitive skills (F = 13.42, p < 0.001) and CELF standard scores. The percentage of the variance in CELF scores accounted for by age at implant was 16%, and by cognitive skills 10%. These two factors accounted for 26% of the variance in CELF standard scores. ANOVA with Tukey pairwise comparison showed that the mean CELF standard score from Group 1 was significantly better than Groups 3, 4, and 5. Group 2 was also significantly better than Groups 4 and 5 (Fig. 6). Cross tabulation and χ 2 analysis supported the hypothesis that a greater percentage of children in Group 1 (58%) obtained CELF standard scores within the normal range (χ 2 = 24.980, df = 4, p < 0.001) than Groups 2, 3, 4, and 5.
Speech Production: DEAP
There were no child participants in Group 5 who had completed the DEAP, and smaller numbers in groups 2, 3, and 4. To examine the impact of age at implant, a two-sample t test was used to compare Group 1 (n = 23) with the remaining Groups 2, 3, and 4 combined (n = 53). Results demonstrated significantly more accurate vowels (t = 2.72, df = 72, p < 0.008), consonants (t = 2.51, df = 58, p < 0.015), phonemes (t = 2.66, df = 64, p < 0.01), clusters (t = 1.96, df = 49, p < 0.05), and significantly higher word intelligibility (t = 2.36, df = 50, p < 0.02) for Group 1 compared with combined Groups 2, 3, and 4 (Fig. 7).
Across a range of speech perception measures, and standardized tests of language and speech production, children who received CIs earlier demonstrated superior communication outcomes compared with children who received CIs at later ages. Data supported the hypothesis that a greater percentage of children who received CIs younger than 12 months would demonstrate language outcomes within the normal range for their hearing peers with typical development. In addition to age at implant, the presence of a co-occurring diagnosis of borderline, or low average cognitive skills, was also a significant factor negatively affecting speech perception, language, and speech production in this cohort.
The finding of a negative relationship between age at implant and speech perception outcomes is consistent with previous studies (5–12,15,16), but the present study affords the first opportunity to demonstrate this finding in a large cohort of children who received CIs younger than 12 months. Speech perception scores, achieved by Group 1 children who received CIs younger than 12 months of age, reached ceiling levels for open-set word and sentence testing in quiet by late primary/early secondary time points. The mean open-set word and phoneme scores obtained by (n = 10) Group 1 child participants were 83.7 (SD 7.5) and 93.7 (SD 3.3), respectively. These scores exceed past reported open-set word scores of 80.5% (SD 12.7) (15) for 21 children, and also word and phoneme scores of 51% and 79% respectively for 27 children who received CIs younger than 12 months (16).
In the present study, the mean open-set sentence score achieved by 26 Group 1 children at school entry was 80.8% (SD 18.5), and, by late primary/early secondary, the mean open-set sentence score for 10 Group 1 children was 96.4%. This exceeds past sentence scores of 87.3% reported for 16 teenagers who had significant residual hearing before proceeding with CIs (68). Means also exceeded LNT, PBK word scores, and HINT-C sentences reported for 117 paediatric implant users who received CIs at 2 to 2.5 years (69).
Across all speech perception measures, cross tabulation and χ 2 demonstrated that Groups 1, 2, and 3 obtained mean scores significantly higher than Groups 4 and 5. These results support the use of CIs under 2 years of age, but not specifically younger than 18 months or younger than 12 months for speech perception. Subsequent analysis of the language data, however, supported the importance of access to hearing within the first year of life.
With regard to language, the present study's findings of a negative relationship between age at implant and optimum language results were consistent with the past literature (6,7,15,21,24–30,34,38). Dettman et al. (7) described language growth rates for children implanted younger than 12 months using the Rossetti Infant-Toddler Language Scales (RI-TLS) checklist, and Ching et al. (34) reported results for PLS standard scores. The present study provides the first evidence of the relationship between access to CIs younger than 12 months and optimum standard scores across three standardized tests, PLS, PPVT, and CELF, and for a longer time course postimplant. A significant percentage of Group 1 children in the present study achieved standard scores within the normal range comparable to their hearing peers. The percentages of Group 1 children within the normal ranges for PLS, PPVT, and CELF were 64%, 81%, and 58% respectively. Cross tabulation and χ 2 demonstrated the percentage of Group 1 children in the normal range was significantly greater than for Groups 2 to 5. Past studies have shown steeper language growth for children implanted before 18 months of age, but on average, children did not reach the normative range (26,28).
Tobey et al. (28) explained the inherent difficulty in isolating the effects of age at implant from length of device experience; “because of their earlier exposure to sound, it is logical to hypothesize children using CIs who are implanted at earlier ages will achieve higher levels of spoken language than later implanted children and will be closer to their peers with normal hearing” (p. 220). It could also be argued that the groups in the present study who were older at implant, had less device experience over all. There was also greater variance in outcomes for children who received CIs later. Not only did more Group 1 children achieve language results within the normal range, but their data suggested greater uniformity. Following Tomblin et al.’ (30) model, more Group 1 children in the present study had less time to develop gaps (delays) in their language, and more time to catch up.
The data from the present study suggest greater benefits to language from earlier access to hearing. Although those in Groups 2 and 3 did derive significant benefit, those in Group 1 consistently achieved the highest mean scores. This supports the concept of a critical period under 3 years of age, consistent with Sharma et al.'s findings (48). For those Groups 4 and 5 children who were implanted after 2 years of age, language results were poorer overall, were less likely to fall within the normative range, and showed greater variance. To add to Tomblin et al.’ (30) model, the present study suggests that the child's brain requires greater reorganisation (to unlearn visual pathways and begin to develop auditory pathways) when access to hearing occurs after 2 years of age.
Numerous children in the present study completed standardized language tests on an annual and semiannual basis. It was not within the scope of this study to examine these rates of language growth, but this will be examined in a subsequent study. By using a slope or trajectory of language, it may be possible to partition out the relative effects of age versus experience.
There are no studies that have reported speech production outcomes from standardized test for a significant number of children implanted under 12 months. In the present study, children in Group 1 achieved 95% (SD 5.3) accuracy of vowels, 83% (SD 13) of consonants, 89% (SD 8.9) of phonemes total, and 64% (SD 25) of clusters. They achieved 63% (SD 23) accuracy of word intelligibility. There were fewer children in this study with speech production results, so the comparisons therefore lacked statistical power. There was evidence, however, to suggest that children in Group 1 had better speech production accuracy than the older children. Further study of speech production (examining accuracy, percentile ranks, and standard scores with a larger group) is planned.
It was not the focus of the present study to describe surgical outcomes, but it can be reported that there were no major surgical complications that could be attributed to a younger age in any of the 403 children under study. Surgical considerations for experienced paediatric teams included recognition of the smaller mastoid size and increased vascular diploic bone. Children under 2 years of age underwent surgery with specialized paediatric anaesthetists. The increased prevalence of acute otitis media and otitis media with effusion in children younger than 12 months did not delay cochlear implant surgery unduly; however, planned simultaneous bilateral CI surgery was occasionally revised to sequential surgery because of a child's middle ear condition (70).
Data collection, from the three paediatric early intervention/cochlear implant centers, with their respective research collaborative partners necessitated a focus on measures that could reasonably be obtained on all children. For this reason, there were a number of other variables, known from the literature to affect child communication outcomes, such as sex, age at hearing-aid fitting, presence of co-occurring disabilities, and other family factors such as maternal education, relative socio-economic advantage, and linguistic opportunities in the home that could not be accounted for in this study. Szagun and Stumper (71), for example, determined that the child's home language environment contributed significantly to the child's opportunities to derive benefit from early implantation. Leigh et al. (21) suggested that parents with greater motivation, who are well informed, may tend to proceed with CIs earlier rather than later. This type of covariance between variables such as maternal education and age at implantation may confound the simplistic view that age at implant is the only factor.
The present study controlled for progressive/acquired hearing loss, and mild/moderate/severe cognitive impairment (by excluding data from these children), but did not specifically examine complex medical conditions and/or aetiologies. Fitzpatrick et al. (72) demonstrated that late implantation could be accounted for by the complexity of determining audiometric levels and parental uncertainty, both of which were exacerbated in child patients with additional special needs.
The present study also did not consider the relative advantages or disadvantages of sign–bilingual, oral–aural, or auditory–verbal approaches that were used with these children. Although debate has raged for decades about the best communication approach to use with children who have significant hearing loss, the fact that communication outcomes are now being reported within the normal range for children using CIs implies that the provision of auditory skills as early as possible in childhood is probably far more important than the communication approach. Future analyses will explore these educational factors with this large cohort. Furthermore, given the importance of vocabulary and oral language in the development of literacy skills, research is currently underway which examines literacy development and academic outcomes for this cohort.
In conclusion, results of the present study of a large cohort of Australian children support the provision of cochlear implants for children with significant hearing loss before 24 months to optimize speech perception, before 12 months to facilitate speech production accuracy, and before 12 months to enable language acquisition to be comparable to hearing peers. Up to 80% of children who received CIs younger than 12 months in the present study demonstrated receptive vocabulary knowledge (PPVT) within the normal range by school entry.
1. Colletti L, Mandala M, Shann RV, et al. Estimated net saving to society from cochlear implantation in infants: A preliminary analysis. Laryngoscope
2. Schroeder L, Petrou S, Kennedy C, et al. The economic costs of congenital bilateral permanent childhood hearing impairment. Pediatrics
3. Forli F, Arslan E, Bellelli S, et al. Systematic review of the literature on the clinical effectiveness of the cochlear implant procedure in paediatric patients. Acta Otorhinolaryngol Ital
4. Luxford WM, Eisenberg LS, Johnson KC, et al. Cochlear implantation in infants younger than 12 months. Int Congr Series
5. Colletti V, Carner M, Miorelli V, et al. Cochlear implantation at under 12 months: Report on 10 patients. Laryngoscope
6. Colletti L, Mandalà M, Zoccante L, et al. Infants versus older children fitted with cochlear implants
: Performance over 10 years. Int J Pediatr Otorhinolaryngol
7. Dettman SJ, Pinder D, Briggs RJ, et al. Communication development in children who receive the cochlear implant younger than 12 months: Risks versus benefits. Ear Hear
8. Valencia DM, Rimell FL, Friedman BJ, et al. Cochlear implantation in infants less than 12 months of age. Int J Pediatr Otorhinolaryngol
9. Holman MA, Carlson ML, Driscoll CLW, et al. Cochlear implantation in children 12 months of age and younger. Otol Neurotol
10. Roland JT, Cosetti M, Wang KH, et al. Cochlear implantation in the very young child: Long-term safety and efficacy. Laryngoscope
11. Waltzman S, Roland JT. Cochlear implantation in children younger than 12 months. Pediatrics
12. Robbins A, Koch DB, Osberger MJ, et al. Effect of age at cochlear implantation on auditory skill development in infants and toddlers. Arch Otolaryngol Head Neck Surg
13. Holt RF, Svirsky MA. An exploratory look at pediatric cochlear implantation: Is earliest always best? Ear Hear
14. Tajudeen BA, Waltzman SB, Jethanamest D, et al. Speech perception
in congenitally deaf children receiving cochlear implants
in the first year of life. Otol Neurotol
15. Wie OB. Language
development in children after receiving bilateral cochlear implants
between 5 and 18 months. Int J Pediatr Otorhinolaryngol
16. Leigh J, Dettman S, Dowell R, et al. Evidence-Based approach for making cochlear implant recommendations for infants with residual hearing. Ear Hear
17. Wie OB, Falkenberg ES, Tvete O, et al. Children with a cochlear implant: Characteristics and determinants of speech recognition, speech recognition growth rate, and speech production
. Int J Audiol
18. Robbins A, Osberger MJ. Meaningful Use of Speech Scale Indiana University School of Medicine, Indianapolis, IN. 1991.
19. Archbold S, Lutman M E, Nikolopoulos T. Categories of auditory performance: Inter-user reliability. Brit J Audiol
20. Moog JS, Geers AE. Epilogue: Major findings, conclusions and implications for deaf education. Ear Hear
21. Leigh J, Dettman S, Dowell R, et al. Communication development in children who receive a cochlear implant by 12 months of age. Otol Neurotol
22. Black J, Hickson L, Black B, et al. Paediatric cochlear implantation: Adverse prognostic factors and trends from a review of 174 cases. Cochlear Implants Int
23. Tsao F, Liu H, Kuhl PK. Speech perception
in infancy predicts language
development in the second year of life: A longitudinal study. Child Dev
24. Fagan MK, Pisoni DB. Hearing experience and receptive vocabulary development in deaf children with cochlear implants
. J Deaf Stud Deaf Educ
25. Nicholas JG, Geers AE. Will they catch up? The role of age at cochlear implantation in the spoken language
development of children with severe to profound hearing loss. J Speech Lang Hear Res
26. Niparko JK, Tobey EA, Thal DJ, et al. Spoken language
development in children following cochlear implantation. JAMA
27. Richter B, Eissele S, Laszig R, et al. Receptive and expressive language
skills of 106 children with a minimum of 2 years’ experience in hearing with a cochlear implant. Int J Pediatr Otorhinolaryngol
28. Tobey EA, Thal D, Niparko JK, et al. Influence of implantation age on school-age language
performance in pediatric cochlear implant users. Int J Audiol
29. Connor C, Craig HK, Raudenbush SW, et al. The age at which young deaf children receive cochlear implants
and their vocabulary and speech-production growth: Is there an added value for early implantation? Ear Hear
30. Tomblin BJ, Barker BA, Spencer LJ, et al. The effect of age at cochlear implant initial stimulation on expressive language
growth in infants and toddlers. J Speech Lang Hear Res
31. Flexer C. Auditory brain development: A paradigm shift for children who are deaf or hard of hearing. Presented at the Australasian Conference on Listening and Spoken Language
, Brisbane, Australia, 2007.
32. Hockett CF. The origin of speech. Sci Am
33. Nicholas JG, Geers AE. Spoken language
benefits of extending cochlear implant candidacy below 12 months of age. Otol Neurotol
34. Ching TY, Dillon H, Day J, et al. Early language
outcomes of children with cochlear implants
: Interim findings on the NAL study on longitudinal outcomes of children with hearing impairment. Cochlear Implants Int
35. Svirsky MA, Teoh SW, Neuburger H. Development of language
and speech perception
in congenitally, profoundly deaf children as a function of age at cochlear implantation. Audiol Neurotol
36. Cuda D, Murria A, Guerzoni L, et al. Pre-school children have better spoken language
when early implanted. Int J Pediatr Otorhinolaryngol
37. Houston DM, Miyamoto RT. Effects of early auditory experience on word learning and speech perception
in deaf children with cochlear implants
: Implications for sensitive periods of language
development. Otol Neurotol
38. May-Mederake B. Early intervention and assessment of speech and language
development in young children with cochlear implants
. Int J Pediatr Otorhinolaryngol
39. Markman TM, Quittner AL, Eisenberg LS, et al. Language
development after cochlear implantation: An epigenetic model. Neurodevelop Disord
40. Ertmer DJ, Jung J. Prelinguistic vocal development in young cochlear implant recipients and typically developing infants: Year 1 of robust hearing experience. J Deaf Stud Deaf Educ
41. Ertmer DJ, Young NM, Nathani S. Profiles of vocal development in young cochlear implant recipients. J Speech Lang Hear Res
42. Schauwers K, Gillis S, Daemers K, et al. Cochlear implantation between 5 and 20 months of age: The onset of babbling and the audiologic outcome. Otol Neurotol
43. Schauwers K, Gillis S, Govaerts PJ. The characteristics of prelexical babbling after cochlear implantation between 5 and 20 months of age. Ear Hear
44. Ertmer DJ, Inniger KJ. Characteristics of the transition to spoken words in two young cochlear implant recipients. J Speech Lang Hear Res
45. Ertmer DJ, Kloiber DT, Jung J, et al. Consonant production accuracy in young cochlear implant recipients: Developmental sound classes and word position effects. Am J Speech-Language Pathol
46. Ertmer DJ, Strong LM, Sadagopan N. Beginning to communicate after cochlear implantation: Oral language
development in a young child. J Speech Lang Hear Res
47. Tobey EA, Geers AE, Brenner C, et al. Factors associated with development of speech production
skills in children implanted by age five. Ear Hear
48. Sharma A, Dorman MF, Spahr AJ. A sensitive period for the development of the central auditory system in children with cochlear implants
: Implications for age at implantation. Ear Hear
49. Geers AE, Lane HS. Central Institute for the Deaf Preschool Performance Scale (CID-PPS). WoodDale, IL: Stoelting; 1984.
50. Griffiths R. (revision by Huntley M 1996). Griffiths Mental Development Scales from Birth to 2 years. Association for Research in Infant and Child Development. Henley-on-Thames, Oxon. 1996.
51. Bayley N. Bayley Scales of Infant Development. 2nd ed.San Antonio, TX: Psychological Corporation; 1993.
52. Wechsler D. Wechsler Intelligence Scale for Children-Revised (WISC R). New York: Psychological Corp; 1994.
53. Wechsler D. Wechsler Preschool and Primary Scale of Intelligence-(Revised). New York: Psychological Corp; 1989.
54. Wechsler D. Wechsler Intelligence Scale for Children-3rd Edition (WISC-III). New York: Psychological Corp; 1992.
55. Wechsler Non Verbal Test Wechsler Non Verbal Test (WNV) Pearson, 2006.
56. Dumont R, Willis JO. Leiter International Performance Scale-Revised. New York: Wiley; 2007.
57. Dettman SJ, Fiket H, Dowell R, et al. Speech perception
results for children using cochlear implants
who have additional special needs. Volta
58. Peterson GE, Lehiste I. Revised CNC lists for auditory tests. J Speech Lang Hear Res
59. Bench J, Bamford J. Speech-Hearing Tests & the Spoken Language
of Hearing-Impaired Children. London: Academic Press; 2002.
60. Zimmerman IL, Steiner VG, Pond RE. Preschool Language Scale
. 4th ed. (PLS-5). Bloomington, MN: Pearson, 2011.
61. Pearson, Zimmerman IL, Steiner VG, Pond RE. Preschool Language
Scale—Fifth Edition (PLS–5). 2011.
62. Dunn LM, Dunn DM. Peabody Picture Vocabulary Test
. 3rd ed. Circle Pines, MN: American Guidance Service, 1997.
63. Dunn LM, Dunn DM. Peabody Picture Vocabulary Test
. 4th ed. Minneapolis, MN: NCS Pearson, Inc, 2007.
64. Wiig EH, Secord W, Semel E. Clinical Evaluation of Language
Fundamentals Preschool—Second Edition, Australian and New Zealand Standardised Edition (CELF P-2 Australian and New Zealand). Toronto, Canada: The Psychological Corporation; 2006.
65. Semel E, Wiig EH. CELF-Fourth Edition, Australian Standardised Edition (CELF-4 Australian). Marrickville: Harcourt Assessment; 2006.
66. Dodd B, Hua Z, Crosbie S, et al. Diagnostic Evaluation of Articulation and Phonology. London, UK: The Psychological Corporation; 2002.
67. Serry T, Blamey P, Spain P, et al. CASALA: Computer aided speech and language
analysis. Australian Commun Quart
68. Dettman SJ, D’Costa WA, Dowell RC, et al. Cochlear implants
for children with significant residual hearing. Arch Otolaryngol Head Neck Surg
69. Eisenberg LS, Kirk KI, Martinez AS, et al. Communication abilities of children with aided residual hearing: Comparison with cochlear implant users. Arch Otolaryngol Head Neck Surg
70. Holland JF, Galvin KL, Briggs RJS. Planned simultaneous bilateral cochlear implant operations: How often do children receive only one implant? Int J Ped Otorhinolaryngol
71. Szagun G, Stumper B. Age or experience? The influence of age at implantation and social and linguistic environment on language
development in children with cochlear implants
. J Speech Lang Hear Res
72. Fitzpatrick EM, Ham J, Whittingham J. Pediatric cochlear implantation: Why do children receive Implants late? Ear Hear
2015; Open Access online.