Contextualization of the Sampling

In 2013 and 2014, of the initial 449 mothers, only 81 could be located to participate in a new phase of the study called Infanto-Álcool I. This phase focused specifically on the audiological evaluation of the present research (Infanto-Álcool II). Although 75 relatives of the 81 mothers were successfully contacted, 6 mothers declined the invitation to participate and 18, despite accepting the invitation, did not show up on the evaluation day even after three attempts; finally, 51 children were evaluated.

Sampling

The 51 children (13–14 years) included both sexes and comprised those who were exposed or not exposed to alcohol during the gestational period, and they underwent the audiological evaluation. Two audiologists with diagnostic expertise (H.O.S. and S.Z.), who were blinded to the alcohol exposure status of each child, performed the tests. The coordinator of the projects identified the children who were exposed to alcohol after the audiological assessment and then divided all children into two groups, namely, the exposed group (EG) with 22 children and the non-EG (NEG) with 23 children. Six children were excluded due to incomplete questionnaires (exclusion criterion). This blind procedure aimed to reduce the occurrence of bias.

Characterization of Alcohol Consumption

Of the 45 children that were included in the present study, 48.9% (22/45) were children of mothers who reported alcohol ingestion at some point during pregnancy. Regarding the frequency of ingestion, 36.4% (8/22) were children of mothers who reported alcohol consumption in the first trimester, 18.9% (4/22) of the children's mothers ingested alcohol in two trimesters, and 45.4% (10/22) of the children's mothers ingested alcohol in three trimesters. In the first trimester, the duration of substance use ranged from 1 to 12 days (mean, 1.9 ± 2.4 days), the number of doses was ranged from 1 to 5 (mean, 1.7 ± 1.0), and the number of times when more than three doses were consumed ranged from 1 to 12 times (mean, 1.8 ± 2.3). In the second trimester, the duration of substance use varied from 1 to 2 days (mean, 1.2 ± 0.5 days), the number of doses was 1–5 (mean, 1.6 ± 1.2), and the number of times when more than three doses were consumed ranged from 1 to 2 times (mean, 1.2 ± 0.5). In the third trimester, the duration of alcohol consumption ranged from 0 to 1 day (mean, 0.1 ± 0.3 days), the number of doses ranged from 0 to 3 (mean, 0.2 ± 0.7), and the number of times when more than three doses were consumed ranged from 0 to 1 (mean 0.1 ± 0.3).

Assessments

None of the children in this research had phenotypes associated with fetal alcohol exposure at 12 years of age when they were assessment by medical team that was trained by the project coordinator [24, 25]. We must acknowledge the limitation of our assessment approach because these diagnostic criteria have a low specificity in detecting alcohol exposure [26]. All the children were evaluated by a psychiatrist under the supervision of one of the authors (E.F.F.) to identify any mental disorder using the Kiddie Schedule for Affective Disorders and Schizophrenia for School-Age Children − Present and Lifetime Version (K-SADS-PL) [27, 28, 29].

Auditory Measurements

Auditory Sensitivity

Only one subject (4.5%, 1/22 in the EG) had altered thresholds with very severe mixed first-degree hearing loss, i.e., unilateral hearing loss in the LE. After otorhinolaryngological evaluation, the patient was diagnosed with labyrinthitis ossificans and was excluded from the other evaluation and analysis steps.

The remaining 44 subjects had hearing thresholds ≤20 dB HL. The comparative study of the airway thresholds between the RE and LE in each group showed that there was no difference for any frequency (EG: 0.25 kHz, p = 1.0; 0.5 kHz, p = 1.0; 1 kHz, p = 1.0; 2 kHz, p = 0.387; 3 kHz, p = 0.113; 4 kHz, p = 1.0; 6 kHz, p = 0.381; and 8 kHz, p = 0.826; NEG: 0.25 kHz, p = 0.613; 0.5 kHz, p = 0.171; 1 kHz, p = 0.307; 2 kHz, p = 1.0; 3 kHz, p = 1.0; 4 kHz, p = 1.0; 6 kHz, p = 0.744; and 8 kHz, p = 0.627). Figure ​Figure11 shows the audiometric pattern, by ear; statistical non-significance allows the groups to be combined.

Open in a separate window

Fig. 1

Mean value comparison of air-conducted thresholds between groups (n = 44 adolescents).

Cochlear Function

The TEOAE examination was performed for all 44 subjects with sensitivity within normal standards, and the presence of responses was observed in both ears. First, the magnitude of the responses between the RE and LE was analyzed (considering the groups separately), and there was no difference in the amplitude values (EG: t = −0.12 and p = 0.781; NEG: t = 0.63 and p = 0.302). This outcome allowed the results of the two ears to be combined for comparative analysis between the two groups. In the NEG, 23 subjects (46 ears) were included, and in the EG, 21 subjects (42 ears) were included. The ± of the response (dB SPL) was 11.22 ± 5.52 dB SPL for the NEG and 11.54 ± 4.93 dB SPL for the EG. The results of the independent sample t test showed no difference between them (t = 0.28; p = 0.776) (Fig. ​(Fig.22).

Open in a separate window

Fig. 2

Average and 95.0% confidence interval of the TOAE amplitude response.

Sample Characterization

For the constitution of the groups, the results of the SQ were adopted because it aims to track the consumption pattern of pregnant women with a “yes” or “no” answer, with the identification of the gestational period. The SQ identified that 48.9% of pregnant women in the present sample consumed some amount of alcohol during pregnancy. This percentage is slightly higher than the estimate in Brazil, which is 20.0–40.6% [36, 37]. In terms of sociodemographic characteristics of the mothers between the two groups, no significant differences were found.

The follow-up of the sample over several years allowed for finding that none of the children had presented, in their development, the set of phenotypic signs traditionally related to fetal alcohol exposure. This characteristic is important for the discussion of the results of auditory tests performed here, as it can be inferred that they presented with less severe effects of alcohol [3, 4]. All adolescents underwent a psychiatric evaluation during “Infanto-Álcool I.”

Among psychiatric diagnoses, attention should be paid to the high occurrence of ADHD, in both the groups. ADHD has an estimated prevalence 2–7% in the infant population [38], but it has been identified in 94.8% of those with FAS using K-SADS [39]. The fact that the participants in the present study did not have signs of FAS may justify the higher occurrence of ADHD than in the general population, but it is lower than that reported in the study by Fryer et al. [39]. Regarding the unexposed group, the occurrence was higher than that in the general population [38], and it was higher than the 30% reported by Fryer et al. [39]. It is necessary to note the possibility of the occurrence of two sources of bias. First, the children who participated in our study already had behaviors that were identified as suspicious by their parents, which led to their inclusion in the study 12 years after the first phase. Second, there are reports regarding the relationship between low socioeconomic status and ADHD [40, 41]. The authors suggest that a lower socioeconomic level increases the risk of ADHD; however, there is a more significant occurrence of psychiatric disorders among family members, people who consume more drugs and tobacco, etc. [41]. There was no difference in family income between the groups, but more than half of the respective families could be classified as low-income families in Brazil (up to five minimum wages [MWs]).

Auditory Sensitivity

The prevalence rate of hearing loss identified in the EG was 4.5%; however, this is lower than that reported in the literature on any hearing loss in children exposed to alcohol during pregnancy, which ranges from 70 to 77% [7, 8, 9], and for sensorineural hearing loss, which is estimated to be 29% [7, 8, 9, 12]. It is necessary to consider the sample composition of the subjects studied here, who had fetal alcohol exposure, but did not have the phenotypic alterations commonly associated with FAS, possibly indicating lower exposure severity. The high frequency of conductive losses owing to external and middle ear alterations has been described in a population with FAS and craniofacial deformities, including, but not limited to, cleft palate, and characteristics that are independent of the conductive hearing loss etiology are risk factors for this type of loss [7, 8, 9, 11, 12]. Sensorineural hearing loss was previously reported to be between 27% and 29% in the presence of FAS [12]. Studies have reported that fetal alcohol exposure can damage the optical placoid of guinea pigs as well as cause deformities in the stereocilia of hair cells and sustentation; fetal alcohol exposure can also cause embryological damage to the inner ear and cranial nerve VIII [8, 13, 14, 42, 43].

However, hearing loss of one adolescent in the EG in this study was identified to be of a mixed type (middle and inner ear impairment), and this is not necessarily representative of the FAS population [7, 8, 9, 12]. The otorhinolaryngological diagnosis for this mixed hearing loss was labyrinthitis ossificans, which is characterized by ossification of the membranous labyrinth, is usually related to a sequel of otological infection, often suppurative labyrinthitis or meningitis, and is not related to ingestion of alcohol during pregnancy [44, 45, 46].

Another result that is contrary to previous reports is that for children without hearing loss in both groups, the thresholds for pure tone were similar. Children with FAS without hearing loss tended to have lower thresholds for frequencies from 3 to 8 kHz (on average 5 dB HL), whereas for frequencies from 9 to 20 kHz, the decrease was greater, from 3 to 14 dB HL [47]. Unfortunately, pure-tone audiometry at high frequencies (>8 kHz) was not possible in the population studied here.

Cochlear Function

Although otoacoustic emission is not an audibility test, the recording of the contractility of outer hair cells, as a result of a sound, provides relevant information on cochlear sensory function [48]. No differences were found in these responses between adolescents with and without fetal alcohol exposure, as has already been reported [47]. Katbamna et al. [47] evaluated eight subjects with FAS in the age range of 8–17 years using distortion product otoacoustic emissions, which indicated a tendency of decreased amplitude in the frequency range of 4–8 kHz. The divergence from the results of the cochlear response pointed out above can be attributed to the type of otoacoustic emissions used. The TEOAE (chosen by us) does not provide the specificity of sound frequency, as that of the respective author mentioned above; they used the distortion product otoacoustic emission method. However, it can also be suggested that the adolescents in this study experienced less severe fetal alcohol exposure, which is compatible with the absence of phenotypic signals.

Assessment of Auditory Processing

The two groups studied showed results suggestive of alteration in at least one auditory processing test, but there was no statistical difference between them. Our results differ from those of other reports [12, 19]. There are some hypotheses for this divergence, such as the differential diagnosis regarding the severity of exposure to alcohol and the occurrence of morbidities in the NEG. The samples evaluated in other studies required a diagnosis of FAS, whereas the adolescents in the present study had not been diagnosed with FAS diagnosis. The second hypothesis concerns the presence of comorbidities in the NEG. The occurrence of psychiatric disorders was the same between the two groups, including ADHD, which is a well-documented condition identified in APD [34, 47]. Thus, both the groups presented risk criteria for altered results of their respective tests [15]. Regardless of whether they were exposed to alcohol during pregnancy, it is still possible to record the overlap between these disorders since children diagnosed with APD also have symptoms of inattention and hyperactivity [18, 49]. The justification for the overlap is the organization of specific neural networks that are shared for more than one activity [50]. However, some studies suggest that APD is more likely to overlap with language impairment and reading disorders than with ADHD [51, 52]. The third and final hypothesis is related to differences in the type of test used and the peripheral hearing condition of the adolescents evaluated between the studies. Church et al. [10] used two tests, SIN and competitive sentences in contralateral (dichotic) modality, and reported that 100% of the subjects had altered results for at least one of the two tests. However, it is important to consider that of the 22 subjects who were recruited, 17 had peripheral hearing loss when auditory processing was evaluated. Of these 22 subjects, 12 underwent behavioral tests. Peripheral hearing loss is not an impediment to auditory processing, but it negatively influences performance [34]. Thus, although the authors mentioned above described a high percentage of altered results, it is not possible to distinguish between acoustic (peripheral) encoding damage and neural processing as the reason for those results. Domellöf et al. (2009) [19] used only one behavioral auditory test, the verbal dichotic with words, in the stages of directed listening. In a dichotic stimulation, if only one of the two stimuli is repeated, those coming from the RE will be the most frequent in right-handed subjects. This performance would reveal the dominance of the left hemisphere for processing verbal language in right-handed subjects [20]. In contrast, ignoring stimuli from the RE and repeating only the stimuli from the LE would imply that the performance of the “top-down” mechanism that is cortical, non-auditory modalities would be acting on the auditory performance [34]. This mechanism was the objective of the study, and the authors showed that right-handed children with FAS had worse directed listening performance than children in a compared to a control group. However, the authors attributed this to developmental disorders or focal brain injury [19]. We also used a dichotic verbal test; however, it is not comparable to the tests used by the authors because we did not perform targeted listening, since we aim to have tests with less influence of top-down mechanisms on auditory tests.