Article information

Korean J Otorhinolaryngol-Head Neck Surg. 2026;.kjorl-hns.2026.00192

Publication date (electronic) : 2026 August 18

doi : https://doi.org/10.3342/kjorl-hns.2026.00192

Sunseong Kang, Hyun Joon Shim, Yong-Hwi An

Department of Otorhinolaryngology-Head and Neck Surgery, Nowon Eulji Medical Center, Eulji University School of Medicine, Seoul, Korea

강순성, 심현준, 안용휘

노원대학교 의과대학 노원을지대학교병원 이비인후과학교실

Address for correspondence Yong-Hwi An, MD, PhD Department of Otorhinolaryngology-Head and Neck Surgery, Nowon Eulji Medical Center, Eulji University School of Medicine, 68 Hangeulbiseok-ro, Nowon-gu, Seoul 01830, Korea Tel +82-2-970-8276 E-mail an0072@daum.net

Received 2026 March 16; Revised 2026 June 18; Accepted 2026 June 25.

Abstract

Introduction

Horizontal canal benign paroxysmal positional vertigo (HC-BPPV) is diagnosed clinically, and the affected side is localized using positional nystagmus tests, including the headroll test, also referred to as the supine-roll test, and lying-down nystagmus/head-bending nystagmus (LDN/HBN). In typical HC-BPPV canalolithiasis, the affected side is the side with stronger nystagmus during the head-roll test. LDN is directed toward the unaffected side, whereas HBN is directed toward the affected side. By contrast, in HC-BPPV cupulolithiasis, the affected side is the side with weaker nystagmus during the head-roll test. LDN is directed toward the affected side, whereas HBN is directed toward the unaffected side [1,2]. In 75%-80% of patients with HC-BPPV, affected-side localization based on the head-roll test is known to agree with localization based on LDN/HBN, whereas localization is discordant in the remaining 15%-20%. In some cases, nystagmus intensity is similar between the left and right sides during the head-roll test, and LDN/HBN may be absent [3,4].

The treatment for HC-BPPV is the barbecue roll maneuver, also known as the Lempert 360-degree roll maneuver. In this maneuver, the patient lies supine, the head is first turned 90 degrees toward the affected side, and the head is then rotated toward the unaffected side in 90-degree steps at 30-60-second intervals, for a total rotation of 270-360 degrees. In HCBPPV canalolithiasis, a single barbecue roll maneuver achieves a treatment success rate of 70%-90%. However, if the head-roll test and LDN/HBN indicate discordant lateralization, selecting the direction of the barbecue roll maneuver becomes difficult, which may lead to a lower treatment success rate. Because of these uncertainties, the American Academy of Otolaryngology-Head and Neck Surgery (AAO-HNS) clinical practice guideline describes the treatment method for HC-BPPV but does not assign a specific recommendation grade. The guideline recommends only the diagnosis of HC-BPPV using the head-roll test. This contrasts with posterior canal BPPV, for which both diagnosis and treatment are recommended in the AAO-HNS BPPV guideline [1].

In this study, we investigated the proportions of patients with HC-BPPV in whom the affected side was concordant or discordant between the head-roll test and LDN/HBN. We also aimed to evaluate whether treatment outcomes differed when the affected side was discordant between the head-roll test and LDN/HBN.

Subjects and Methods

We retrospectively reviewed the medical records of 490 patients who visited our hospital between November 2021 and December 2023 with positional dizziness and had no notable findings other than HC-BPPV on history taking and physical examination. All patients underwent otoendoscopy and video nystagmography, including assessment of spontaneous nystagmus, gaze-evoked nystagmus, head-shaking nystagmus, and positional nystagmus. Positional nystagmus was assessed in the following order: HBN/LDN, the Dix-Hallpike test, and the head-roll test. During each positional nystagmus test, the position was maintained for at least 1 minute. LDN was observed after the patient was placed in the supine position, and HBN was observed with the head bent forward. For LDN/HBN, nystagmus with a slow-phase velocity of ≥1°/s was recorded together with its direction.

Patients were excluded if the affected side could be clearly attributed to other inner ear diseases, such as sudden sensorineural hearing loss, vestibular neuritis, or Ménière’s disease. Patients were also excluded if they showed persistent geotropic direction-changing positional nystagmus, as in light cupula, if the affected side could not be distinguished because nystagmus intensity was similar bilaterally on the head-roll test, if LDN and HBN showed the same nystagmus direction, if spontaneous nystagmus was present, if unilateral canal paresis was 25% or greater on the caloric test, or if they had traumatic or multicanal BPPV. A total of 283 patients were included in the final analysis, with a mean age of 61.6±13.2 years, an age range of 22-87 years, and a male-to-female ratio of 83:200.

Among the 283 patients diagnosed with HC-BPPV on the head-roll test, 139 patients were classified as having geotropic positional nystagmus and 144 as having apogeotropic positional nystagmus. Based on the LDN/HBN findings, patients were further classified into three groups: the concordant group, in which the affected side was the same as that determined by the head-roll test, the discordant group, in which the affected side differed from that determined by the head-roll test, and the control group, in which LDN/HBN was not observed. In all patients, treatment was performed using the barbecue maneuver based on the affected side identified by the head-roll test. In patients with apogeotropic positional nystagmus, a vibrator was applied to the mastoid region on the affected side for 30 seconds before the barbecue maneuver. Treatment outcomes were evaluated based on symptoms and nystagmus findings at presentation and 2 weeks later. If nystagmus remained on positional nystagmus testing after each treatment, the barbecue maneuver was performed one additional time. Treatment success was defined as the complete resolution of both dizziness symptoms and nystagmus.

All statistical analyses were performed using SPSS version 12.0 (SPSS Inc.). Variables were expressed as mean±standard deviation. Age was compared among the three groups using one-way analysis of variance, and sex ratio and treatment success rate were compared among the three groups using the chi-square test. A p value of <0.05 was considered statistically significant. This study was approved by the Institutional Review Board at Nowon Eulji Medical Center (approval number: EMCS-2021-03-020).

Results

Among the 283 patients, the interval from symptom onset to the outpatient visit was 5.08±7.14 days (0-30 days). The mean number of treatment sessions was 1.44±0.77 (1-4 sessions). The affected side was left in 172 patients (60.8%) and right in 111 patients (39.2%).

Among the 139 patients with HC-BPPV who showed geotropic positional nystagmus on the head-roll test, LDN or HBN was observed in 69 patients. Of these, both LDN and HBN were observed in 41 patients, LDN alone in 10 patients, and HBN alone in 18 patients. These patients were classified into the concordant or discordant group according to the side indicated by each nystagmus pattern and the side indicated by the head-roll test. Among all 139 patients with geotropic HC-BPPV, 40 patients (28.8%) showed concordant affected-side localization on the head-roll test and LDN/HBN and were assigned to the concordant group, whereas 29 patients (20.9%) showed discordant localization and were assigned to the discordant group. In the remaining 70 patients (50.3%), LDN/HBN was not observed and these patients were assigned to the control group (Table 1).

Table 1.

Demographic characteristics of geotropic subgroup and resolution rate after 2 weeks

Among the 144 patients who showed apogeotropic positional nystagmus on the head-roll test, LDN or HBN was observed in 65 patients. Of these, both LDN and HBN were observed in 42 patients, LDN alone in 14 patients, and HBN alone in 9 patients. These patients were classified into the concordant or discordant group according to the side indicated by each nystagmus pattern and the side indicated by the head-roll test. Among all 144 patients with apogeotropic HC-BPPV, 47 patients (32.6%) showed concordant affected-side localization and were assigned to the concordant group, whereas 18 patients (12.5%) showed discordant localization and were assigned to the discordant group. In the remaining 79 patients (54.9%), LDN/HBN was not observed and these patients were assigned to the control group (Table 2).

Table 2.

Demographic characteristics of apogeotropic subgroup and resolution rate after 2 weeks

At the 2-week outpatient follow-up, both symptoms and nystagmus had resolved in 104 of the 139 patients with geotropic BPPV, and the treatment success rate for geotropic HC-BPPV was 74.8%. Among the 144 patients with apogeotropic positional nystagmus, both symptoms and nystagmus had resolved in 83 patients, and the treatment success rate for apogeotropic HC-BPPV was 57.6%. In the geotropic HC-BPPV group, the treatment success rates were 72.5% in the concordant group, 68.9% in the discordant group, and 78.6% in the control group. There was no significant difference in treatment success rate among the three groups (Table 1). In the apogeotropic HC-BPPV group, the treatment success rates were 51.1% in the concordant group, 50.0% in the discordant group, and 63.2% in the control group. The treatment success rate did not differ between the concordant and discordant groups, whereas the control group, in which LDN/HBN was not observed, showed the highest treatment success rate, and the difference among the three groups was statistically significant (Table 2).

Discussion

BPPV has an annual incidence of approximately 10.7-64 cases per 100000 population. Its incidence increases sharply after the age of 40 years [2], and it occurs most commonly in individuals in their 50s and 60s. Idiopathic BPPV is particularly frequent in women, with a frequency approximately 2-3 times higher than that in men [2]. In the present study, the mean age of the patients was 61.6±13.2 years, and the male-to-female ratio was 83:200. These findings appear to reflect demographic characteristics consistent with previous reports. In addition, considering previous reports in which nystagmus could not be analyzed because of preceding vestibulopathy, we excluded secondary causes, including Ménière’s disease, head trauma, vestibular neuritis, sudden sensorineural hearing loss, and otitis media, and analyzed treatment outcomes in idiopathic HC-BPPV [5]. The mean interval from symptom onset to outpatient presentation was 5.08±7.14 days, and the mean number of treatments was 1.44±0.77. HC-BPPV was first described by McClure in 1985 [6], and the 2017 clinical practice guideline of the AAO-HNS recommends the head-roll test for diagnosis [1]. However, in actual clinical practice, bilateral symmetric nystagmus may be elicited during the head-roll test, making it difficult to determine the affected side. To overcome this limitation, various adjunctive positional tests have been proposed.

Choung, et al. [7] attempted to localize the affected side using the bow and lean test, and the concepts of LDN and HBN were subsequently introduced [3,4,7]. LDN/HBN is currently used together with the head-roll test as an adjunctive method for affected-side localization, and several studies have reported its diagnostic utility. Koo, et al. [3] reported that the direction of LDN was concordant with the head-roll test in 77% of cases with cupulolithiasis. Lee, et al. [4] reported that the concordance rate of HBN reached 88% in canalolithiasis. In contrast, other studies have reported discordant or absent LDN/HBN findings in 22.2%-44% of cases [1,3-5,8).

In the present study, an affected side discordant with that indicated by the head-roll test was observed in 21% of the geotropic HC-BPPV group and 12% of the apogeotropic group. This discrepancy may be explained by several factors. Koo, et al. [3] reported that nystagmus observed during supine head-position changes does not always conform to Ewald’s second law. This may be attributable to latent nystagmus, positional nystagmus observed even in healthy adults, and differences in the initial location of otoliths. In addition, biomechanical variables, including the angle and velocity of head rotation during the head-roll test, otolith location, examiner experience, and patient responsiveness, may collectively contribute to different nystagmus patterns even in the same lesion [3,9,10)]. Oh, et al. [5] noted that the HC is tilted approximately 30° anteriorly in the normal upright position. They also suggested that LDN may not be observed if otoliths are located in the lower portion of the non-ampullary arm. Accordingly, in the present study, LDN was assessed immediately after HBN without returning the patient to the sitting neutral position. Lee, et al. [4] reported that, in their study, 2 of 32 patients changed from canalolithiasis to cupulolithiasis during HBN. This suggests that the subtype may change during positional nystagmus testing and that caution is needed when interpreting these tests.

The absence of a significant difference in treatment success rate between the concordant and discordant groups in the present study is likely related to the natural course of HCBPPV. Imai, et al. [11] reported that spontaneous remission occurs within several days to several months in approximately 70%-80% of patients with HC-BPPV, with a mean remission period of approximately 16±19 days. Shim, et al. [12] reported that the mean duration of dizziness symptoms was 6.7 days in geotropic HC-BPPV and 3.7 days in apogeotropic HC-BPPV, indicating a relatively short natural course. In addition, Kim, et al. [8] performed the first follow-up at 2-3 days and reported an overall remission rate of 65%. Given that the follow-up time point in the present study was set at 2 weeks after treatment, spontaneous remission may have partly contributed to the observed treatment effect. Future studies assessing short-term treatment outcomes immediately after treatment or within several days may provide a more accurate conclusion. The similar treatment success rates observed when treatment was performed according to the localization direction of the head-roll test may suggest that head-roll test findings were more closely aligned with treatment outcomes than LDN/HBN findings in this cohort. However, because treatment was not performed according to the LDN/HBN results in the present study, direct comparison was not possible.

In the apogeotropic group, the higher treatment success rate in the control group, in which LDN/HBN was not observed, suggests that disease severity may have been low enough for nystagmus not to be detected. Although this could not be definitively demonstrated because nystagmus intensity was not quantitatively measured in the present study, Lee, et al. [9] reported that, among patients who underwent the bow and lean test, the maximum slow-phase eye velocity on the head-roll test was lower in those without nystagmus than in those with nystagmus. Based on this finding, the absence of LDN/HBN may indicate a relatively smaller size or amount of otolith debris, resulting in a more favorable prognosis. In the geotropic group, the lack of a significant difference in treatment success rate among the three groups may be explained by the generally favorable treatment response in this subtype. However, because the sample size of the discordant group was small in the present study, the low statistical power should be considered when interpreting the results.

Compared with previous studies reporting that LDN/HBN was absent in 20%-50% of cases, the proportion of patients in the control group in the present study was approximately 50%-55%. Previous studies suggested that LDN/HBN may have disappeared because otoliths returned to their original position during the head-roll test or because otoliths located on the utricular side in apogeotropic nystagmus became dislodged during testing. However, these mechanisms do not explain the present findings, because the head-roll test was performed after LDN/HBN had been assessed. In the present study, cases with very weak LDN/HBN, in which the presence or absence of nystagmus was difficult to determine, may have been classified into the control group, leading to a higher proportion of control-group patients. Alternatively, the reproducibility and clinical utility of LDN/HBN in routine clinical practice may be lower than previously reported.

Kim, et al. [13] reported that light cupula was observed in approximately 14% of patients with persistent geotropic direction-changing positional nystagmus. Although such pathophysiologic characteristics may have contributed to variability in diagnosis and treatment response in some cases, the present study excluded patients suspected of having light cupula and included only cases of idiopathic HC-BPPV.

This study has several limitations. First, because of its retrospective design, only the treatment response according to the head-roll test findings could be analyzed, and detailed analysis of treatment response according to LDN/HBN findings was limited. Second, the sample size of the discordant group was small, resulting in low statistical power and limiting the generalizability of the results. Third, because treatment effects were evaluated at 2 weeks, spontaneous remission may have confounded the observed treatment effect, making it impossible to compare short-term treatment outcomes.

Future studies should include follow-up immediately after treatment on the same day and within several days to exclude spontaneous remission, and treatment effects should be evaluated more clearly by confirming the number of treatment sessions. In addition, the effect of selecting the treatment direction according to LDN/HBN findings on treatment success rate should be analyzed, and large-scale prospective randomized controlled trials including patients in whom the affected side cannot be determined on the head-roll test are needed [14,15]. Furthermore, comparative studies of different therapeutic maneuvers, including the barbecue roll, Gufoni, Appiani, and Forced prolonged position maneuvers, should be performed together with quantitative analyses of nystagmus intensity, duration, and latency. Branco, et al. [16] reported that ampullopetal flow-induced nystagmus during the Gufoni and Gufoni-Appiani maneuvers was associated with a lower improvement rate. This suggests that observing nystagmus during the barbecue roll maneuver may help predict treatment outcomes. Finally, detailed analyses of pathophysiologic differences according to anatomical location, including the ampullary arm, non-ampullary arm, and cupular attachment site, are expected to contribute to improving the diagnostic accuracy and treatment efficiency of HC-BPPV.

In this study, patients with HC-BPPV showed similar treatment outcomes when the affected side was discordant between the head-roll test and LDN/HBN and when the affected side was concordant between the two tests. Therefore, when the affected side cannot be accurately determined using the head-roll test in HC-BPPV, LDN/HBN-based lesion localization should be interpreted with caution. Because potential error due to the small sample size cannot be excluded, future prospective controlled studies with larger patient populations are needed to compare treatment outcomes more accurately.

Supplementary Materials

Article information Continued