To the Editor:

One of the main factors of sudden hearing impairment, vestibular disturbance, or tinnitus is generally thought to be an acute labyrinthine ischemia of varying degrees. Despite the most common mechanism of sudden hearing loss appearing to be impaired cochlear blood circulation,1 the scientific basis of this assumption has not yet been proven. Experimental studies performed in animals through ferromagnetic obstruction of cochlear blood vessels have shown that vascular impairment of the inner ear results in a considerable decrease in intracochlear oxygenation, causing a significant loss in the auditory response.2 In humans, deafness associated with small infarctions of cochlear tissue have exceptionally been reported in the context of rare multifocal microangiopathic encephalopathies,3 but no data are available in the field of current sudden hearing impairment.

Our attention was stimulated by some of our patients with thromboembolic disease who spontaneously reported unilateral acute perception deafness. We thus performed a retrospective case-control study.

Three hundred sixty-eight patients, corresponding to 101 men and 267 women, median age 41 years (range, 17 to 72), with antecedents of spontaneous deep vein thrombosis (thrombosis group: TG), were investigated over a 3-year period in the outpatient Department of Hematology. Eighteen of them, 12 women and 6 men, 38 to 69 years old, had also suffered from acute unilateral hearing impairment. Explorations were in favor of a pure unexplained perception auditory defect, with computed tomography scanning and magnetic resonance imaging showing no cerebral lesion. These 18 patients had no antecedents of symptomatic atherosclerosis, except the oldest man who had angina pectoris, and ultrasound exams performed on asymptomatic patients gave no evidence of atherosclerotic lesions on carotid arteries. None of the women had hypertension and seven of them were smokers: the three youngest, aged 38, 43, and 45 years, were also oral-contraceptive takers. Three of the men were heavy smokers, including the one with angina pectoris; two of them had mild hypertension.

During the same period of time, acute unilateral perception deafness could only be found in 4 of 395 nonthrombotic consecutive patients studied for hemorrhagic symptoms or thrombocytopenia (hemorrhage group: HG), and in 6 of the 395 nonthrombotic and nonhemorrhagic sex- and age-matched control individuals (control group: CG) who had a systematic medical check-up in a Public Health Center (Pearson chi-square test: P = .0008). No difference could be evidenced between the hemorrhage group and the control group.

Concerning biological risk factors for thrombosis, the prothrombin 20210A allele was positive in 25 patients from TG (6.8%), in 4 patients from HG (1%), and in 3 patients from CG (0.7%) (P = .0008). The factor V gene G1691A mutation (factor V Leiden) was positive in 63 patients from TG (61 heterozygous and 2 homozygous; 17.1%), in 5 from HG (1.3%), and in 6 from CG (1.5%;P < .0001). Antithrombin deficiency was evidenced in 5 patients from TG (1.4%), but in none from HG or CG (P = .005). Protein C deficiency was found in 11 patients from TG (3%), in 2 patients from HG (0.5%), and in 3 patients from CG (0.8%; P = .006). Protein S deficiency was confirmed in 8 patients from TG (2.2%), in 1 patient from HG (0.25%), and in 1 patient from CG (P = .005). Antiphospholipid antibodies could be found in 35 patients from TG (9.5%), in 6 patients from HG (1.5%), and in 7 patients from CG (1.8%; P < .0001). Concerning the type of positive antiphopholipid antibodies, a lupus anticoagulant was found in 22 patients from TG but in only 4 patients from HG and 5 patients from CG (P < .001). Twenty-seven patients from TG had positive anticardiolipin IgG antibodies whereas only 3 patients from HG and 4 patients from CG were positive (P < .001). Finally, 23 patients from TG were positive for anti–β2-glycoprotein I IgG antibodies, but only 3 patients from HG and 2 from CG (P < .0001). The prevalence of positive biological risk factors for thrombosis was the same in the hemorrhage group and in the control group.

Due to the absence of differences in terms of prevalence of deafness and of biological risk factors for thrombosis, we defined a normal group as the addition of the hemorrhage group and of the control group. Taking into account this reference group and the thrombosis group, we calculated crude odds ratios (cOR) for perception deafness and 95% confidence intervals (95%CI) using logistic regression for each biological venous thrombosis risk factor and for thromboembolic antecedents (Table 1). Thereafter, a multivariate logistic regression was performed to adjust to potential confounding factors (adjusted odds ratio, aOR; a stepwise procedure entering each variable with P lower than .20 in monovariate logistic regression; table 1): only a positive thromboembolic antecedent and a positive prothrombin gene 20210 A allele were confirmed to be independent risk factors for perception deafness.

Table 1.

Odds Ratios and 95%CI for Perception Deafness and Type of Positive Biological Risk Factor for Thrombosis in Thrombotic Patients, With the Normal Group as Reference: Results of the Univariate and of the Multivariate Logistic Regressions

Among the 18 thrombotic patients with deafness, 6 were heterozygous carriers of the prothrombin gene 20210A allele. Among these 6 carriers were the 3 young smoking and oral-contraceptive–taking women, another nonsmoking woman and two smoking men, including the oldest one with angina pectoris. The introduction of gender, oral contraceptives, and smoking as new parameters for logistic regression analysis gave nonsignificant results, showing that these parameters are not independent risk factors for perception deafness: results concerning smoking are concordant with available published data.4 However, in young thrombotic women with the 20210A allele in the prothrombin gene, we cannot exclude that smoking and oral contraceptive use may be clinical cofactors of the disease.

In conclusion, these data seem to indicate that, in patients with thromboembolic antecedents, the prothrombin 20210A allele is an independent risk factor for perception deafness. This may constitute a new indirect argument for the vascular/thrombotic origin of some of these accidents. As a consequence, patients with thrombotic antecedents and the prothrombin 20210A allele should be preserved from aggressive conditions for the internal ear (excessive noise, scuba diving, treatments with a known specific toxicity, etc). If patients are young women, as they may be putative clinical cofactors, hearing preservation may be an additional argument for oral contraceptive and smoking cessation. The prothrombin 20210 allele has already been described to increase the risk of myocardial infarction among non-Mediterranean young women and the relative risk was higher in smoking carriers.5 We have previously described cases of spinal cord infarction in young smoking and oral-contraceptive–taking women with the prothrombin 20210A allele.6 The 20210A allele of the prothrombin gene may thus have a more general status of risk factor for arteriolar thrombosis, a generally difficult-to-diagnose and sometimes silent event. Some anatomical localizations, due to their dramatic consequences, are however evident. Future questions should address the identification of the spectrum of additional abnormalities acting to increase the penetrance of clinical arteriolar thrombotic manifestations associated with the prothrombin 20210A allele.