Abstract
Among patterns of hearing loss, sudden sensorineural hearing loss (SSNHL) ranks as one of the least common, accounting for 1% of all cases of hearing loss with approximately 4000 new cases reported each year (1). However, the inability to prepare for the sudden loss of function, the lack of a satisfying explanation for its cause, and the limited treatment options also make it a particularly devastating disorder. A specific cause for the hearing loss is found in less than 15% of cases and guidelines for management of SSNHL have only recently been published (2,3).
Prior association studies have shown that traditional vascular risk factors such as smoking, hypertension, dyslipidemia, and diabetes can be risk factors for SSNHL, with one study even showing a higher risk of subsequent stroke in patients with SSNHL (4). However, meta-analyses have shown that these associations are also inconsistent (5). Although a history of migraine is not generally considered to be a risk factor for SSNHL, it has also not been systematically investigated. Case reports have indicated that some patients who experience SSNHL also experience other symptoms attributable to migraine and that sudden hearing loss associated with a severe migraine headache can be associated with ischemic changes in the inner ear (6,7).
In this volume, Chu et al. report on the association between migraine and the incidence of SSNHL over a 10-year period using the National Health Insurance Research Database in Taiwan, which is based on a universal health care system with centralized information on all diagnostic codes and prescriptions. The study was designed as a retrospective case-control study, with the incidence of SSNHL in migraine patients and controls in the year 2000 used as the baseline. Migraine and SSNHL were both diagnosed by ICD-9-CM codes, as were major vascular risk factors such as diabetes, hypertension, dyslipidemia, and atrial fibrillation. After exclusion of a history of SSNHL at baseline, Meniere’s disease, and acoustic neuromas, 10,280 migraine patients were compared to 41,120 controls. These subjects were followed for a median of five years until one of three outcomes was obtained: 1) development of SSNHL, 2) subject death, or 3) end of the study in 2009. The study required that the diagnosis of migraine be coded by a neurologist and that the diagnosis of SSNHL be coded by an otolaryngologist.
The authors report a 1.8-fold increased risk of SSNHL in patients with migraine with a very small but detectable cumulative risk of SSNHL each year throughout the 10 years of observation. This cumulative risk amounted to 0.4% in migraine patients and 0.2% in controls. In a multivariate analysis, comorbidity with hypertension increased the risk to 1.92 in migraine patients, but this was not statistically significant. The role of hypertension in the risk of SSNHL in controls was not reported, however.
The study provides a global perspective on the incidence of SSNHL. The authors calculated an incidence of 81.6 cases in 100,000 person-years in migraine subjects and 45.7 cases per 100,000 person-years in controls. The relative risk increased to 118.6 per 100,000 person-years in migraine subjects over the age of 40 years. These estimates are much higher than the often-reported incidence of five to 20 cases in 100,000 person-years for SSNHL (2,3,5). However, some widely cited incidence data in the current literature are extrapolated from as few as 18 patients cared for in one health care system or are reported from personal communications (8,9). Older studies also did not use the current, more generally accepted criteria for SSNHL that require a loss of 30 dB of hearing in three consecutive frequencies occurring over less than 72 hours. Conversely, the reported incidence in the study was lower than the 160 per 100,000 person-years reported in 2004 from a German population (10).
One challenge to making an accurate assessment of the incidence of SSNHL in any setting is that acuity can often be difficult to determine, especially if the patient recovers quickly. Between one-third and two-thirds of patients will recover some hearing within two weeks, potentially leaving many milder cases undiagnosed (11). Furthermore, since SSNHL is usually associated with tinnitus and often associated with vertigo, some practitioners may code these cases as labyrinthitis or Meniere’s disease. The initial symptoms of SSNHL can also be extremely vague, with the patient simply complaining of an ear feeling full or “plugged.” After ruling out the more mundane causes of subacute hearing loss such as cerumen impaction and middle ear effusion (which cause conductive hearing loss), the physician must maintain a high index of suspicion for SSNHL as early therapy with oral or intratympanic steroids may be beneficial (3). In reality, the diagnosis of SSNHL is often not certain until enough clinical follow-up or diagnostic testing has been obtained. In one early study, an initial diagnosis of idiopathic SSNHL of 76 patients was changed in 24 cases in follow-up, with the subjects being diagnosed with other inner ear disorders such as Meniere’s disease, labyrinthitis, and acoustic neuromas (9). These are issues that would challenge incidence reports in any setting but diagnostic accuracy in large-scale studies is more likely to be compromised, a sacrifice that is made in favor of larger numbers.
A further issue with diagnostic coding data is that initial wrong coding usually cannot be reversed and then only with extreme effort. In clinical practice, diagnoses are often made without following recommended research criteria and sometimes a code is chosen because it is the closest approximation to the choice the clinician wants, but is not quite the right diagnosis. It is unknown how closely the otolaryngologists had followed the criteria for SSNHL reported by the study authors. We would expect for this to be a problem equally prevalent in cases as well as in controls in this study. However, given the association between migraine and Meniere’s disease, it may be helpful to know in this study whether migraine patients who were diagnosed with SSNHL were more likely than controls to subsequently incur a diagnosis of Meniere’s disease. The study reports that 227 migraine patients were excluded because of a diagnosis of Meniere’s disease. How does this compare to the number of Meniere’s diagnoses in controls?
For the purposes of this study, any subject with a diagnosis of Meniere’s disease was excluded from the analysis. However, legitimate SSNHL and subsequent Meniere’s disease can coexist. Damage to the inner ear from any cause can subsequently lead to a process called “delayed endolymphatic hydrops,” which behaves very similarly to idiopathic endolymphatic hydrops (i.e. classic Meniere’s disease) except that it is typically associated with more profound preexisting hearing loss in the symptomatic ear. It would still be coded as Meniere’s disease as there is no specific code for delayed endolymphatic hydrops. Therefore, it would be helpful to know how many subjects in the migraine and control groups were excluded because the diagnosis was changed from SSNHL to Meniere’s disease, or whether a subsequent diagnosis of Meniere’s disease was added. In the latter case, it would have been inappropriate to exclude the subject.
One curious finding of the study is the overall low rate of migraine with aura in the migraine group and an even lower proportion of migraine subjects with SSNHL being diagnosed with aura. Of the 43 patients with migraine diagnosed with SSNHL, only three had migraine with aura. Though the relative risk of developing SSNHL in this group was 2.28, it was not statistically significant. This would argue against Viirre and Baloh’s suggestion that SSNHL has an etiology similar to migraine aura with a common mechanism of vasospasm. Viirre and Baloh presented 13 cases of SSNHL occurring in migraine patients, six of whom had also experienced visual aura (6). In five of the six subjects who experienced unilateral migraine headaches, the hearing loss was on the same side as the headache, suggesting a common lateralized mechanism to both headache and hearing loss. However, in Viirre and Baloh’s study, the subjects were well-characterized patients of the investigators who were chosen for their unique history of combined SSNHL and migraine symptoms and thus there may have been a selection bias. In the study by Chu et al., the diagnosis of migraine was made by a neurologist, a design that may have selected for patients with more clinically active migraine and perhaps a more reliable diagnosis but that may have also led to patients with inactive migraine symptoms being considered as controls.
Because pathological studies of the inner ear after SSNHL are rare, our understanding of the mechanism of SSNHL will likely continue to be based on the “company it keeps” for some time. Migraine may indeed be an important risk factor in certain individuals, perhaps those who also have vascular risk factors like hypertension. Whether it is an important risk factor in the larger scheme of SSNHL given the variety of other potential causes remains to be validated. However, Chu et al.’s well-motivated study is a first investigation of an uncommon but clinically very significant otological disorder presented in the context of much more prevalent public health problems.