The study group consisted of 346 consecutive patients, aged 21.8±19.9 years (0.16–84 years), including 176 males and 170 females, with sensorineural HL of 30dB or more (calculated as the average hearing threshold at frequencies relevant for speech – 0.5, 1, 2, and 4kHz) and on whom genetic tests were done to identify mutations in GJB2 and GJB6. The HL was confirmed by pure tone audiometry or, in the case of children less than 6 years old, Brainstem Evoked Response Audiometry (BERA). Among the subjects studied, 91 had cochlear implants.

DNA was extracted from the peripheral blood by a salting-out procedure.13 The typing for the c.35delG mutation in GJB2 gene was performed by two independent methods: AS-PCR and PCR-RFLP, as described previously.14,15 Briefly, multiplex amplification of three GJB2 regions (c.12-72, c.68-198 and c.306-464) with fluorescent labeled primers was performed. In parallel, PCR-RFLP analysis specific for c.35delG mutation was conducted. Size of the fragments generated by the two above-mentioned protocols was determined using a 3500XL Genetic Analyzer (Life Technologies) with Gene Mapper (Life Technologies) software. Samples lacking two c.35delG mutations were further analyzed by direct DNA sequencing of the PCR amplified exon 2 of the gene. Sequence analysis was performed with Variant Reporter v1.1 (Life Technologies) software. Genotyping of the c.-23+1G>A GJB2 mutation was performed using real-time PCR according to the manufacturer's instructions (Life Technologies), while the GJB6 mutation del (GJB6-D13S1830) was detected using the method of del Castillo et al.16 Genetic nomenclature is in line with recommendations of the Human Genome Variation Society (HGVS).

All 346 patients underwent a 12 lead Electrocardiogram (ECG). ECG recording was performed and evaluated using the Sentinel Cardiology Information Management System (Reynolds Medical). The durations of PR interval, QRS complex, and QTc interval were measured by a single observer on a 5 times enlarged image on a computer monitor with a cursor accuracy of 1ms (sampling frequency 1000Hz). PR interval was usually measured in lead II, QRS complex in V1 and/or V5, and QT interval in at least three leads of the same cycle (usually leads II, V2, and V5). Heart rate was calculated from the RR interval preceding the cycle of the QT interval measured. To correct QT interval due to heart rate (i.e., QTc), Bazett's formula was used for rates between 50 and 120bpm; for faster or slower rates Hodges’ formula was used. In order to calculate average QTc, patients with QRS complexes ≥120ms were excluded. A QTc interval above 460ms in women, 450ms in men, and 440ms in children under 14 years of age, was diagnosed as prolonged.

The study population was divided into three groups based on reference values of QTc for children, women, and men,17 giving 147 children <14 years (42.4% of patients), 100 women ≥14 years (28.9%), and 99 men ≥14 years (28.6%). The group of children included 70 (47%) girls.

Statistical analysis of each of the three groups involved a Kruskal–Wallis test and a Wilcoxon two-sample test.

The study was approved by the Bioethics Committee, approval number KB/05/2009. The study was done in accordance with the principles of the Declaration of Helsinki. All patients or their guardians gave written consent to participate.

Mutations in GJB2 were identified in 112 (32%) patients with HL. No mutations were detected in the GJB2 and GJB6 genes analyzed in the other 234 (68%).

The mean age of children with a GJB2 mutation and children without GJB2 mutation was similar (6.0±4.6 vs. 5.8±4.5 years, p=0.86). The mean age of adults in both groups (with and without GJB2 mutation) was also similar (35.1±17.6 vs. 30.5±16.5 years, p=0.17 for women; 31.6±18.9 and 30.3±17.5 years, p=0.99 for men) (Table 1).

The homozygous genotype c.[35delG];[(35delG)] in the GJB2 gene was detected in 72 (20.8%) subjects. In 40 (11.8%) individuals, heterozygous GJB2 genotypes were identified: c.[35delG];[=] (n=15), c.[35delG(;)313_326del] (n=5), c.[35delG(;)-23+1G>A] (n=4), c.[313_326del];[=] (n=2), c.[35delG(;)551G>C] (n=2), c.[35delG(;)95G>A] (n=2), and a compound GJB2 and GJB6 heterozygous genotype [GJB2:c.35delG];[GJB6:del(GJB6-D13S1830)] (n=2). The following GJB2 heterozygous genotypes occurred in individual cases: c.[35delG(;)269T>C], c.[148G>A];[=], c.[79G>A; 341G>A; 269T>C], c.[101T>C];[=], c.[35delG(;)235delC], c.[35delG(;)358_360delGAG], c.[35delG(;)139G>T], c.[334_335delAA];[=] (Table 2).

The next phase of the study was to compare electrocardiographic parameters in the study population. The measured parameters were compared between two groups: the group of patients with a mutation in gene GJB2 (both homozygotes and heterozygotes) and the group of patients with no mutation in the gene. No significant differences were found for any measured parameter in any of the groups studied. Table 3 presents inter-group differences in PR interval; Table 4 presents inter-group differences in QRS complex; and Table 5 presents inter-group differences in QTc interval.

In addition, we compared parameters in patients with profound HL caused by a GJB2 mutation homozygous genotype c.[35delG];[(35delG)] and in those with a heterozygous genotype [GJB2:c.35delG];[GJB6:del(GJB6-D13S1830)] versus all other patients with other types of HL. The analysis revealed no significant differences (Table 6).

Finally, we analyzed the incidence of ECG abnormalities in the group of patients with a mutation in GJB2 and those without any mutation in this gene. The abnormalities in the group of patients with a mutation were: atrial rhythm (2 patients), sinus bradycardia (n=12), short time atrioventricular conduction (n=3), left anterior fascicular block (n=1), a QTc interval above the norm for age and sex (n=10), and signs of myocardial inferolateral wall necrosis associated with prolonged QTc (n=1). In the group of patients without any mutation, abnormalities were: sinus bradycardia (n=25), atrial rhythm (n=3), wandering atrial pacemaker (n=1), short time atrioventricular conduction (n=2), first degree atrioventricular block (n=5), incomplete right bundle branch block (n=1), left anterior fascicular block (n=1), complete bundle branch block (n=7), intraventricular conduction disturbances (n=3), signs of left ventricular hypertrophy (n=3), signs of myocardial ischemia (n=2), and prolonged QTc (n=23). In one boy, there was supraventricular rhythm with a prolonged atrioventricular conduction time, nonspecific intraventricular conduction disturbances, and signs of right ventricular hypertrophy.

In the group of patients with a GJB2 mutation and prolonged QTc (n=10), the mean QTc interval was 463.2±15.6ms, whereas in patients with prolonged QTc and without any mutation (n=23), the mean QTc interval was 460.3±12.8ms. This difference was also not significant (p=0.69).

Our study did contain some limitations. Different phenotypic expressions are the result of various GJB2 mutations. Since the described GJB2 gene mutations that result in HL are recessive, we compared parameters from a group of patients with a homozygous c.[35delG];[(35delG)] genotype and a heterozygous [GJB2:c.35delG];[GJB6:del(GJB6-D13S1830)] genotype with all other patients. The results of such an analysis are limited by the small number of people with a homozygous c.[35delG];[(35delG)] genotype, and there were only 2 children with heterozygous [GJB2:c.35delG];[GJB6:del(GJB6-D13S1830)] genotype giving the phenotypic effect of deafness.

In the study group, the presence of heart defects has not been ruled out by a reliable test (e.g. echocardiography), and neither has the effect of certain medications which might have influenced electrocardiographic parameters. The study group was selected by screening of people with HL who did not report symptoms of heart disease or did not take drugs, producing a group of apparently healthy people who had HL.