The authors have declared that no competing interests exist.
Conceived and designed the experiments: CR MW. Performed the experiments: DW AB CR MW. Analyzed the data: DW KG AB MW BN. Contributed reagents/materials/analysis tools: CR MW. Wrote the paper: MW KG BN.
The pathomechanism and location of idiopathic sudden sensorineural hearing loss (ISSHL) is unclear. In a previous case-control study, we found elevated fibrinogen concentrations and a higher prevalence of T allele carriers of the glycoprotein (Gp) Ia C807T polymorphism in ISSHL patients.
127 patients with ISSHL (mean age 53.3 years, 48.8% females), who underwent a standard therapy with high dose steroids, pentoxifyllin and sterofundine over 8 days were included. We examined the influence of GpIa genotype and fibrinogen (BclI-, A312-, HaeIII-) genotype and fibrinogen plasma levels on hearing recovery after 8 weeks (change from baseline: 0 dB = no recovery, >0 to 10 dB = moderate recovery, >10 dB = good recovery). In a subsample of 59 patients with ISSHL, we further studied the association of platelet glycoprotein GpIa, Ib and IIIa densities on hearing recovery as well as the possible effect-modification of platelet glycoproteins on hearing recovery by plasma fibrinogen.
In univariate analysis, neither the GpIa genotype nor fibrinogen genotype (all p>0.1) but lower fibrinogen levels (p = 0.029), less vertigo (p = 0.002) and lower GpIIIa receptor density (p = 0.037, n = 59) were associated with hearing recovery. In multivariate analysis, fibrinogen significantly modified the effect of GPIa receptor density on good hearing recovery (effect-modification on multiplicative scale OR = 0.45 (95% confidence interval (0.21–0.94)), p = 0.03). GPIb receptor density below the mean was associated with a 2-fold increase in good hearing recovery both in patients with fibrinogen levels above (p = 0.04) as well as in patients with fibrinogen levels below the mean (p = 0.06). There was no indication for an effect-modification (p = 0.97).
The findings suggest a vascular/rheological origin of ISSHL with unique features of thrombosis in the inner ear artery that may include complex interrelationships among platelet glycoproteins and plasma fibrinogen.
Sudden sensorineural hearing loss (SSHL) is a dysfunction of the inner ear characterized by sudden onset and rapid progression of hearing impairment within hours or days. It can occur at every age, but mostly affects elderly people and has an estimated incidence of 10-20/100,000/year
Hearing recovery in ISSHL varies widely from complete, spontaneous recovery to non-recovery and shows an unpredictable course
One hypothetical scenario in ISSHL is a sudden reduction of the blood flow in the labyrinthic artery, which is a functional end artery. The blood flow in the labyrinthic artery is regulated mainly by adrenergic receptors on smooth muscle cells. It is also influenced by plasma viscosity and local regulatory mechanisms, but not by central regulatory mechanisms as in larger vessels. Due to the possible ischemia-like cause of ISSHL, it was hypothesized that risk factors, comparable to those associated with ischaemic or thromboembolic diseases, such as heart disease, stroke and thrombosis
We have recently shown that fibrinogen concentrations are significantly elevated in ISSHL patients
Initially, 142 consecutive German patients with a pantonal, sudden, sensorineural hearing loss (ISSHL) were included in the study: 75 male and 67 female patients aged between 14 and 83 years. Patients have previously been described
Variable | All patients; n = 127 | Patients stratified for glycoprotein receptor analysis | |||
With analysis; n = 59 (46.5%) | Without analysis; n = 68 (53.5%) | p | |||
Basic patients' characteristics | Age in years, mean (sd) | 53.3 (17.1) | 59.5 (13.6) | 48.0 (18.1) | |
Female gender, n (%) | 62 (48.8) | 27 (45.8) | 35 (51.5) | 0.521 | |
Smoker, n (%)& | 68 (54.8) | 32 (56.1) | 36 (53.7) | 0.788 | |
Tinnitus, n (%)&& | 91 (75.2) | 39 (72.2) | 52 (77.6) | 0.495 | |
Vertigo, n (%)&& | 46 (37.7) | 18 (32.7) | 28 (41.8) | 0.304 | |
Laboratory values | GPIa-gene | ||||
CC (wild type), n (%) | 28 (22.1) | 9 (15.3) | 19 (27.9) | 0.085 | |
TC or TT, n (%) | 99 (77.9) | 50 (84.7) | 49 (72.1) | ||
BcLI-gene | |||||
CC (wild type), n (%) | 82 (64.6) | 40 (67.8) | 42 (61.8) | 0.478 | |
TC or TT, n (%) | 45 (35.4) | 19 (32.2) | 26 (38.2) | ||
A312-gene | |||||
CC, n (%) | 77 (60.6) | 33 (55.9) | 44 (64.7) | 0.313 | |
TC or TT, n (%) | 50 (39.4) | 26 (44.1) | 24 (35.3) | ||
HaeIII-gene | |||||
CC (wild type), n (%) | 78 (61.4) | 37 (62.7) | 41 (60.3) | 0.780 | |
TC or TT, n (%) | 49 (58.6) | 22 (37.3) | 27 (39.7) | ||
Fibrinogen in mg/dl, mean (sd) | 280.1 (78.9) | 277.5 (72.7) | 282 (84.5) | 0.729 | |
Hearing recovery | None, n (%) | 43 (33.9) | 18 (30.5) | 25 (36.8) | |
1-10 dB, n (%) | 46 (36.2) | 19 (32.2) | 27 (39.7) | 0.160§ | |
> 10 dB, n (%) | 38 (29.9) | 22 (37.3) | 16 (23.5) |
& = data missing on 3 patients; && = data missing on 6 patients; CC = wild type, TC = heterozygous; TT = homozygous; § = Mantel-Haenszel Chi-square test.
Blood was drawn by venipuncture at the time when hearing recovery was determined, which was at least 8 weeks after ISSHL to avoid acute phase reactions secondary to the acute event. Fibrinogen was measured in citrated plasma after centrifugation for 15 min at 2,500 g. The sample analysis was done within 1 h after venipuncture. DNA was extracted from EDTA-treated blood by a spin column procedure (Quiagen, Germany).
Fibrinogen was determined using the Multifibren U reagent (BCS analyzer; Dade Behring). The coagulation process was initiated by adding a large excess of thrombin to the plasma sample. The time until formation of a fibrin clot was measured. The fibrinogen concentration was then calculated from a standard curve prepared by measuring fibrinogen calibrators of a known concentration. All other routine measurements were performed as previously described
The platelet-surface GPIa receptor density (integrin α2β1) was determined by flow cytometry analysis using the antibody originally sold by American Diagnostica, which can now be obtained from Biocytex (Marseille, France). Measurements were performed as described by the company (manual reference number 7008). In order to maintain platelet integrity, utmost care was exercised to avoid platelet activation during collection. The anticoagulant used was trisodium citrate 0.129 M (using a ratio 9∶1 volumes). Blood samples were treated within 4 hours after collection. Blood was stored at room temperature before testing (18–25°C). The test was performed on platelet rich plasma (PRP). 25 µL of PRP was diluted and then added to 20 µL of anti GpI, GpIb or GpIIIa mouse antibody. Tubes were gently vortexed for 1 to 2 seconds, secondary polyclonal anti-mouse IgG-FITC antibody was added and incubated at room temperature for 10 minutes, and finally diluted with 2 ml of dilution solution (obtained from the test kit) before measurements were performed within 2 hours of obtaining the samples. Before cytometric analysis on a FACS-STAR (Coulter) cytometer the PRP antibody mixture was stored at 2–8°C. For calibration analysis, beads coated with increasing and accurately known quantities of mouse IgG were stained in parallel with the sample using the same secondary IgG-FITC antibody and incubated for the same amount of time as per the manufacturer's recommendations. The bead population was gated using a discriminator on forward scatter to minimize background. Mean fluorescence intensity values were interpolated on the calibration curve and the number of molecules read off directly. The GpIa, GpIb and GpIIIa platelet-surface receptor densities were determined in parallel in the same experimental setting. According to the manufacturer, the values for negative isotypic controls were very low (∼300 sites per platelet), so that these controls were not included in the test kit. The number of glycoprotein molecules per platelet, as described by the manufacturer for a normal population, was: 53,000±12,000 for GpIIIa, 38,000±11,000 for GpIb, and 5,000±2,800 for GpIa. In the experiments described here, the number of glycoprotein molecules per platelet was in the same range, except for the case of GPIa, in which 6 participants receptor density markedly exceeded this range (>7,000 molecules per platelet). Therefore, these 6 outliers were removed and data on glycoprotein density analysed for the 53 remaining patients.
The fibrinogen polymorphisms BcII, Aα312, HaeIII and Dusart that may influence fibrinogen concentration or function
fibrinogen
Forward primer
Reverse primer
Probe 1 (labeled with the fluorescent dye VIC)
Probe 2 (labeled with the fluorescent dye FAM)
fibrinogen
Forward primer
Reverse primer
Probe 1 (labeled with the fluorescent dye VIC)
Probe 2 (labeled with the fluorescent dye FAM)
fibrinogen
Forward primer
Reverse primer
Probe 1 (labeled with the fluorescent dye VIC)
Probe 2 (labeled with the fluorescent dye FAM)
fibrinogen
Forward primer
Reverse primer
Probe 1 (labeled with the fluorescent dye VIC)
Probe 2 (labeled with the fluorescent dye FAM)
Platelet glycoprotein Ia C807T
Determination of the C or T allele at position 807 of the GPIa gene was performed by mutagenically separated PCR. We determined the C807T genotype indirectly by identifying the linked G873A genotype. The three following primers were used:
Primer 873A in 5` of exon:
Primer 873G in 3` of exon:
Primer 1 in 5` of exon:
The DNA fragments were visualized on a 3 or 5% agarose gel after staining with ethidium bromide by ultraviolet transillumination.
The study protocol was reviewed and approved by the ethics committee of the University of Münster, Germany. Written informed consent was obtained from all patients. If participants were minors, written informed consent was obtained from next of kin, caretakers, or guardians.
Descriptive statistics encompassed the mean (standard deviation) for normally distributed metric variables and the median (range) for not-normally distributed metric variables, respectively. To test for differences between 2 independent subgroups, we used student's t-test for normally distributed metric variables and Mann-Whitney-U-Test for not normally distributed, metric variables. To test for differences between >2 independent ordered groups (of hearing recovery), we used the Mantel-Haenszel Chi-square test in binary and ordinal variables, and ANOVA for normally distributed metric variables respectively Kruskal-Wallis-Test for not-normally distributed metric variables. Binary logistic regression was used to examine the independent association of fibrinogen plasma level, glycoprotein receptor densities and their interaction with hearing recovery >10 dB (averaged over 4 frequencies, 250 Hz, 1,000 Hz, 2,000 HZ, and 4,000 Hz) versus hearing recovery ≤10 dB. We assumed a positive effect of low levels of both fibrinogen and glycoprotein receptors on hearing recovery
The C807T polymorphism in GpIa is a silent polymorphism: it does not alter the deduced amino acid sequence of the translated protein. However, an association between the expression levels of integrin α2β1 on platelets and the C807T polymorphism has been suggested
The concentration of platelet integrin α2β1 was measured using a highly sensitive fluorescence-based method (
A single color flow cytometric analysis of the platelet glycoproteins GpIa, GpIb and GpIIIa was used (Biocytex, Marseille, France). The number of antigenic sites is determined by converting the fluorescence intensity into the corresponding number of sites per platelet based on calibrated bead standards (a). In all experiments, more than 90% of the platelets of the PRP were gated for the experiment, as shown here for GpIa (b).
The values are shown as boxplots, separated for T allele carriers (phenotype TT or CT) and CC allele carriers. The boundary of the box closest to zero indicates the 25th percentile, the line within the box marks the median, the square indicates the mean, and the boundary of the box farthest from zero indicates the 75th percentile. Whiskers above and below the box indicate the 90th and 10th percentiles. The points represent the 1st and 99th percentiles. P indicates the level of significance according to the Mann-Whitney-U-test. *, P<0.05; ***, P<0.001.
We also measured the platelet expression of the glycoproteins Ib (GpIb), as shown in
As shown in
Variable | All patients stratified for hearing recovery | ||||
None | 1–10 dB | > 10 dB | p | ||
n = 43 (33.9%) | n = 46 (36.2%) | n = 38 (29.9%) | |||
Basic patients' characteristics | Age in years, mean (sd) | 53.4 (16.8) | 52.8 (19.1) | 53.8 (15.3) | 0.966# |
Female gender, n (%) | 25 (58.1) | 21 (45.7) | 16 (42.1) | 0.146§ | |
Smoker, n (%)& | 21 (48.8) | 24 (54.6) | 23 (62.2) | 0.236§ | |
Tinnitus, n (%)&& | 30 (75.0) | 32 (72.7) | 29 (78.4) | 0.741§ | |
Vertigo, n (%)&& | 24 (58.5) | 13 (29.6) | 9 (24.3) | ||
Laboratory values | GPIa-gene | ||||
CC (wild type), n (%) | 29 (67.4) | 31 (67.4) | 30 (79.0) | 0.268§ | |
TC or TT, n (%) | 14 (32.6) | 15 (23.6) | 8 (21.1) | ||
BcLI-gene | |||||
CC (wild type), n (%) | 27 (62.8) | 29 (63.0) | 26 (68.4) | 0.606§ | |
TC or TT, n (%) | 16 (37.2) | 17 (37.0) | 12 (31.6) | ||
A312-gene | |||||
CC (wild type), n (%) | 28 (65.1) | 27 (58.7) | 22 (57.9) | 0.501§ | |
TC or TT, n (%) | 15 (34.9) | 19 (41.3) | 16 (42.1) | ||
HaeIII-gene | |||||
CC (wild type), n (%) | 26 (60.5) | 26 (56.5) | 26 (68.4) | 0.485§ | |
TC or TT, n (%) | 17 (39.5) | 20 (43.5) | 12 (31.6) | ||
Fibrinogen in mg/dl, mean (sd) | 306.0 (91.5) | 266.6 (62.2) | 267.2 (76.3) | ||
Receptor density | GPIa, mean (SD), n = 53$$ | 4,162.9 (1,398.1) | 4,929.9 (1,104.1) | 4,590.6 (940.2) | 0.169 |
(n = 59/53) | GPIb, mean (SD) | 43,604.1 (5,099.6) | 45,947.0 (5,361.6) | 43,300.1 (5,118.5) | 0.239 |
GPIIIa, mean (SD) | 56,658.1 (5,433.4) | 58,976.9 (6,377.1) | 54,190.3 (5,516.4) |
Sd = standard deviation; # = Kruskal-Walis-Test; & = data missing on 3 patients; && = data missing on 6 patients; § = Mantel-Haenszel Chi-square test; CC = wild type, TC = heterozygous; TT = homozygous; $ = ANOVA, $$ = 6 outlier removed.
As shown in
GPIa receptor density above mean | GPIa receptor density below mean | ORs (95%-CI) for GPIa receptor density within different strata of fibrinogen | |||
N with/without hearing improvement | OR (95% CI) | N with/without hearing improvement | OR (95% CI) | ||
Fibrinogen above mean | 1/9 | 1.0 | 5/10 | 1.24 (0.59 – 2.60); p = 0.568 | 1.24 (0.59 – 2.60); p = 0.568 |
Fibrinogen below mean | 8/13 | 1.50 (0.71 – 3.14); p = 0.286 | 5/17 | 0.84 (0.29 – 2.42); p = 0.741 | 0.56 (0.25 – 1.24); p = 0.150 |
Measure of effect modification on additive scale: relative excess risk due to interaction, RERI (95% CI) = −0.90 (−1.97 – 0.17)
Measure of effect modification on multiplicative scale: ratio of ORs (95% CI) = 0.45 (0.21 – 0.94), p =
ORs are adjusted for smoking and vertigo, # = 4 datasets missing due to missing covariates.
GPIb receptor density above mean | GPIb receptor density below mean | ORs (95%-CI) for GPIb receptor density within different strata of fibrinogen | |||
N with/without hearing improvement | OR (95% CI) | N with/without hearing improvement | OR (95% CI) | ||
Fibrinogen above mean | 3/14 | 1.0 | 4/9 | 2.00 (1.04 – 3.86); p = |
2.00 (1.04 – 3.86); p = |
Fibrinogen below mean | 4/15 | 1.28 (0.68 – 2.41); p = 0.453 | 10/17 | 2.52 (0.96 – 6.64); p = 0.061 | 1.98 (0.88–4.43); p = 0.098 |
Measure of effect modification on additive scale: relative excess risk due to interaction, RERI (95% CI) = 0.24 (−1.46 – 1.95)
Measure of effect modification on multiplicative scale: ratio of ORs (95% CI) = 0.99 (0.52– 1.86), p = 0.968
ORs are adjusted for smoking and vertigo, # = 4 datasets missing due to missing covariates.
GPIIIa receptor density above mean | GPIIIa receptor density below mean | ORs (95%-CI) for GPIIIa receptor density within different strata of fibrinogen | |||
N with/without hearing improvement | OR (95% CI) | N with/without hearing improvement | OR (95% CI) | ||
Fibrinogen above mean | 2/11 | 1.0 | 5/12 | 1.57 (0.81 – 3.03); p = 0.178 | 1.57 (0.81 – 3.03); p = 0.178 |
Fibrinogen below mean | 5/17 | 1.44 (0.76 – 2.70); p = 0.260 | 9/15 | 2.62 (0.96 – 7.15); p = 0.060 | 1.82 (0.83 – 4.01); p = 0.135 |
Measure of effect modification on additive scale: relative excess risk due to interaction, RERI (95% CI) = 0.61 (−1.11 – 2.34)
Measure of effect modification on multiplicative scale: ratio of ORs (95% CI) = 1.16 (0.61– 2.20), p = 0.650
ORs are adjusted for smoking and vertigo, # = 4 datasets missing due to missing covariates.
The C807T polymorphism has previously been shown to be a risk factor for incidence of ISSHL
Fibrinogen level (p = 0.029), vertigo (p = 0.002) and lower GpIIIa receptor density (p = 0.037) varied between groups with different hearing recovery.
We found a significant association between lower plasma fibrinogen levels and good hearing recovery. However, there was no significant association of fibrinogen polymorphisms (that may influence fibrinogen concentration or function) with hearing recovery, suggesting that this effect might have resulted from factors other than genetics. Suckfüll et al showed that fibrinogen apheresis improves the outcome in ISSHL patients
Fibrinogen can be influenced by various non-genetic factors, which include: smoking, obesity, diabetes, hormones, and acute inflammation or trauma
Hearing impairment in ISSHL is often accompanied by vertigo
The possible positive effect of low platelet concentrations of GPIIIa on hearing recovery in ISSHL patients has been shown for the first time in this study. The Glycoprotein IIb/IIIa complex (also known as integrin αIIbβ3) is a receptor for fibrinogen and aids in platelet activation. It is a target of several well-established drugs including abciximab, eptifibatide and tirofiban, which are used in treating patients who have unstable angina, certain types of heart attacks, and are used in combination with angioplasty with or without stent placement. Further studies are recommended to prove possible beneficial effects of these drugs in ISSHL. Additionally, because ADP induces exposure of fibrinogen-binding determinants on GpIIb/IIIa complexes, ADP antagonists such as clopidogrel, ticlopidine or prasugrel may have positive effects in ISSHL.
After adjustment for smoking and vertigo, lower levels of all three platelet glycoproteins were associated with better hearing recovery, although this association was significant solely for GPIb.
GpIb, as component of the GpIb/V/IX complex, binds von Willebrand factor and allows platelet adhesion and plug formation at sites of vascular injury. However, there have been conflicting reports on a possible association of polymorphisms of this glycoprotein with myocardial infarction and coronary artery disease
Fibrinogen significantly modified the effect of GPIa receptor density on hearing recovery, but in an unexpected manner: a putative protective effect of low GpIa concentrations, after adjustment for vertigo and smoking, was significantly diminished in patients with lower fibrinogen levels. There was a harmful effect modification by fibrinogen on a multiplicative scale. The origin of this effect is unclear. Apart from complex interactions between fibrinogen and GpIa
Limitations of our study are the relatively small number of patients for whom glycoprotein receptors were measured and the relative high percentage of smokers, both of which may have influenced and biased results. The power of this study was not sufficient to allow for sensitivity analyses in subgroups (e.g. in smokers/non-smokers) or for more complex analyses of interactions between genetics, environmental factors and platelet glycoproteins. Another meaningful supplement to our analysis would have been evaluation of hearing recovery in speech audiometry. Other investigators were able to show that speech audiometry might be more sensitive in detecting differences in hearing recovery
Altogether, data from this study further suggest a vascular/rheological origin of ISSHL with unique features of thrombosis in the inner ear artery that may include a complex interrelationship among platelet glycoproteins and plasma fibrinogen. Further studies with prospective design are necessary to confirm and deepen the findings described here and to prove the proposed therapeutic options, particularly glycoprotein blockade.
The expert technical assistance of Margit Käse is gratefully acknowledged. We thank Dr. Helmut Schulte (PROCAM) for his help with the manuscript preparation and Dr. Sabine Ziemer for critical reading of the manuscript.