A Careful Interpretation of the Audiometry may contribute to the Diagnosis of Alport Syndrome: A Pilot Study

Despite the different pathways of embryonic development, kidneys and ears share similar mechanisms that can influence both organogenesis1–3. Alport syndrome (sAlport), for example, is known for its phenotype of renal, auditory and ocular involvement4 and affects 1/10.000 individuals. The pathophysiology of this disease lay on a genetic defect in type IV collagen synthesis that interferes with the glomerular filtration barrier4 and degenerates cochlear vascular stria5.


Introduction
Despite the different pathways of embryonic development, kidneys and ears share similar mechanisms that can influence both organogenesis [1][2][3] . Alport syndrome (sAlport), for example, is known for its phenotype of renal, auditory and ocular involvement 4 and affects 1/10.000 individuals. The pathophysiology of this disease lay on a genetic defect in type IV collagen synthesis that interferes with the glomerular filtration barrier 4 and degenerates cochlear vascular stria 5 .
Usually, the sAlport presents with sensorineural hearing loss 6 and renal loss of function 4 . The early onset -young adulthood 6 or at childhood 7 -and the family history of renal/hearing impairment may suggest sAlport. Depending on the variants, some individuals present only with renal failure, hardening sAlport diagnosis 8,9 .
The World Health Organization (WHO) suggests that the best ear mean of thresholds (500, 1000, 2000 and 4000 Hz) should not exceed 25 dB. If so, the patient has hearing loss 10 . In many studies the criteria for hearing loss remain unclear 9,[11][12][13][14][15] , others use WHO's recommendations 8 . Both cases can "hide" milder hearing losses.
Because of some incongruencies at setting a cut-off for hearing threshold, we compared the prevalence of hearing loss through different criteria 10,16,17 in two groups of patients with glomerulopathies software, version 26. We present summarized results through relative and absolute frequencies (nominal variables) and mean and standard deviation (numerical variables). Inferential analyses were performed using Pearson's Chi-square test and Wilcoxon's test. We also compared Clark [10] and WHO [12] criteria regarding sensitivity and specificity (ROC curve). We set alpha= 0.05 and 95% confidence interval.
The prevalence of hearing loss varied according to the criteria: no hearing loss (WHO), similar prevalence (GBD; 16.67% in the sAlport group and 14.29% in the FSGS group), and different prevalence (Clark; 66.67% in the sAlport group and 28.57% in the FSGS group) ( Table 1). In sAlport group, two individuals presented mild hearing loss: a man and a woman. Figure 1 shows the thresholds, between 250 and 8000 Hz, on the right and left ears of each group.
-one group with sAlport and the other, the control group, with segmental and focal glomerulosclerosis (FSGS), which share many clinical characteristics.

Materials And Methods
This is a cross-sectional, single-center, pilot study developed at the Outpatient Clinic of Glomerulonephritis from Universidade Federal de São Paulo (UNIFESP), Brazil. This protocol was approved by the Research Ethics Committee of UNIFESP (CAAE 96087918.9.0000.5505) and developed according to the principles of the Helsinki Declaration. All participants provided their written consent.
We included patients younger than 65 years-old that had biopsy-proven kidney disease (optical microscopy, immunofluorescence and, electron microscopy in sAlport). The individuals of the sAlport group presented familial hematuria and typical renal lesion of Alport disease at electron microscopy. Kidney biopsy also confirmed FSGS in the control group. Patients who had other possible causes of hearing loss (exposure to occupational noise, history of repeated otitis media or otosclerosis and type B tympanometry) were excluded.
Demographic and clinical data (weight, height, BMI, gender, age, ethnicity, family history of hearing loss and/or renal disease, high blood pressure, serum creatinine, 24hour proteinuria and presence of hematuria) were obtained from the patients' charts. Audiometry was performed by an audiologist and included external auditory canal examination and tonal and vocal audiometry.
We used the quadritonal mean (Mq) of the auditory thresholds for the frequencies of 0.5, 1, 2 and 4 kHz to determine hearing loss according to three criteria: WHO (Mq > 25 dB) 10 , Global Burden Disease (Mq > 20 dB) 17 and Clark (Mq > 15 dB) 16 .
The data were organized in Microsoft Excel for Mac, version 16.35 and analyzed in IBM® SPSS Statistics Clark's criteria (Mq > 15 dB) detected hearing loss with a higher sensitivity than the WHO (Mq > 25 dB). Our comparison (ROC curve) showed statistical significance between the area-under-the-curve of them (AUC difference = 0.095; p = 0.293, 95% CI -0.07; 0.26) ( Table 2).

Discussion
This is a pilot study that investigated the prevalence of sensorineural hearing loss in patients with sAlport, according to three different criteria: Clark 16 , GBD 17 and WHO 10 . We also compared the sensitivity and specificity of Clark and WHO criteria and observed that Clark's criteria (Mq > 15 dB) favors hearing loss diagnosis. sAlport phenotype remains well known because of its triad: eye, ear and kidney impairment. Their organogenesis seems to be unrelated, but some transcription factors are common to renal and auditory development 1-3 and the clinical presentation lays on the structural defect of type IV collagen. More commonly, sAlport is linked to chromosome X (COL4A5 mutation) and 80% of the individuals have this pattern of inheritance. The most frequent phenotype includes a family history of hematuria, CKD, and earlyonset hearing impairment 7 . The diagnosis of renal  (Clark, 1981) 0.595 0.095 0.293 (-0.07; 0.26) Hearing loss (WHO, 1991) 0.500 Table 2. Comparison between sensitivity and specificity of different hearing loss criteria by the quadratic mean.
Abbreviations: WHO, World Health Organization; AUC, area-under-the-curve; CI, confidence interval. involvement in the Alport Syndrome is based on electron microscopy typical findings. Hearing loss is described as sensorioneural, bilateral, and symmetrical, even though its degree can vary, which we had already observed 13 .
Even for those patients who have isolated high frequency hearing loss (30% to 50%), renal impairment is always present. The opposite is not observed, and renal impairment can occur solely. Also, the severity of ear and kidney dysfunction seemed to be unrelated 18 .
In this study, we used a similar renal groupcorticoresistant familial FSGS -to compare its hearing thresholds to the sAlport's. About 20% of familial FSGS cases present type IV collagen mutations 19 , which is found in all cases of FSGS we used here.
As we recruited individuals from an outpatient service of glomerulonephritis, none of the subjects were on dialysis at the time we conducted this protocol.
FSGS and sAlport groups were comparable in BMI and gender, but not in age because sAlport group was considerably younger than the FSGS group, possibly because the patients were recruited from an outpatient clinic that assists individuals over 12 years old. But, despite their age, sAlport group presented a higher prevalence of hearing loss by Clark's and GBD's criteria, which was unlikely considering their age. They, also, presented a higher frequency of family history for CKD and hearing loss, which were expected. We avoided the confounding factor of age on hearing loss by not including individuals older than 65 years old, which may be affected by presbycusis. Even younger, our sAlport group did not meet the WHO's criteria for hearing loss 7 , which may be found among individuals who are older than 40 years old. Additionally, the Alport Syndrome patients we included did not complain of any clinical hearing loss.
The WHO's criteria (Mq > 25 dB) is widely used in Audiology and is based on expert opinion to suggest its cut-off. More recently, they suggested that a functional loss should be considered in the borderline cases of hearing loss (around 20 dB of quadritonal mean) 10,16,17,20 . With our patients, we observed that GBD's criteria 17 fits better in identifying milder grades of hearing loss, as already shown 21 . Clark's criteria 16 , instead, differentiated the two groups we compared. At Clark's cut-off, patients may not complain about hearing, but they may be aware of difficulties in listening-in-noise situations, which highlights a mild hearing impairment. Despite hearing difficulties, these patients can complain about hearing discomfort or tinnitus, which can worsen their quality of life 22 . Here, we observed that a lower cut-off facilitates the hearing assessment when it comes to sAlport investigation.
The main limitations of our study are the small sample size and the number of patients included in the FSGS Group.
Although sample size limits our statistical analyses, it does not invalidate the discussion of the hearing loss criteria we propose in this pilot study. Also, our study highlights the highly variable presentations of pure-tone audiometry in Alport Syndrome. The absence of hearing complaints made it difficult for us to expand the FSGS group, because these patients were unmotivated to participate in a hearing investigation, which they considered unnecessary.

Conclusion
Depending on the criteria, audiometry can hide mild hearing losses. We consider that a critical evaluation of the hearing thresholds may help physicians to early detect minimal hearing impairment even though the report says, "normal hearing", especially in a group of patients without end-stage renal disease.