Introduction

Amblyopia, commonly referred to as “lazy eye,” is a leading cause of preventable visual impairment in children and affects millions of people worldwide. The pooled prevalence rate of amblyopia was estimated to be between 1.44% and 1.75%, with approximately 99.2 million individuals affected globally in 2019. This number is projected to increase to 175.2 million by 2030 and 221.9 million by 20401, 2. Typically diagnosed in childhood, amblyopia usually results from interruptions in normal visual development, such as strabismus, differences in refractive error between the eyes, or visual deprivation3.

While amblyopia is traditionally seen as a condition limited to the visual system, new research suggests that its effects may extend beyond sight, impacting other neurological functions4. Amblyopia usually arises from disrupted visual experiences during a critical period of brain development, leading to irreversible changes in neural processing that can last a lifetime. Amblyopia, though primarily a cortical visual disorder, may influence neural pathways beyond the visual cortex. Evidence suggests that abnormal visual input during early development affects multisensory integration, including visuo-vestibular and visuo-auditory networks. Neuroimaging studies have demonstrated altered functional and structural connectivity in amblyopic individuals, involving not only the occipital lobe but also parietal and cerebellar regions involved in spatial orientation and postural control5,6,7. Such disruptions may impair the central processing required for vestibular compensation, potentially increasing vulnerability to disorders like Benign paroxysmal positional vertigo (BPPV).

One intriguing yet less explored area is the potential connection between amblyopia and vestibular dysfunction. BPPV is one of the most common causes of peripheral vertigo, characterized by brief episodes of dizziness or spinning sensation induced by changes in head position8. BPPV is most often attributed to displaced otoliths—calcium carbonate crystals—within the semicircular canals of the inner ear, which can disrupt normal balance signals sent to the brain9. The lifetime prevalence of BPPV is approximately 2.4%, corresponding to 11 to 64 cases per 100,000 population. The 1-year prevalence is around 1.6%. BPPV is most commonly observed in women between the ages of 45 and 709, 10. Although BPPV is a benign condition, it can significantly affect daily activities and quality of life, especially in older adults who face higher risks of falls and related injuries11. Aside from age-related changes in vestibular function, some studies suggest that central nervous system factors, including sensory integration and processing, may play a role in BPPV’s development12. Given this context, we hypothesized that individuals with amblyopia—who may exhibit altered multisensory integration—would have a higher incidence of BPPV than those without amblyopia.

In this study, we aim to investigate whether amblyopia is linked to a higher risk of developing BPPV, using data from the National Health Insurance Service-National Sample Cohort (NHIS-NSC), a large, population-based database. By examining the risk of BPPV in individuals with amblyopia compared to matched controls, we seek to determine if this developmental visual impairment is indeed associated with an increased likelihood of vestibular dysfunction. Such findings could shed light on the broader neurological impacts of amblyopia and highlight the importance of integrated vision care from an early age.

Materials and methods

Database and type of study

This study used the National Health Insurance Service - National Sample Cohort (NHIS-NSC) database. This is a database of 1 million individuals constructed by random sampling 2% of a total of 50 million people in the Republic of Korea, and is a representative sample cohort with evenly distributed sex, age, economic status, and residential areas. This database includes demographic data with diagnostic code based on the International Classification of Diseases, 10th revision (ICD-10), treatment history, patients’ visit date, and medications.

Study population and case definition

Amblyopia is defined as reduced best-corrected visual acuity in one or both eyes without a definite organic cause. We identified individuals diagnosed with amblyopia between 2002 and 2004 using the ICD-10 code H530. To improve diagnostic specificity, a positive amblyopia diagnosis must satisfy all the following conditions: (1) the patient was diagnosed with amblyopia under ICD-10 code H530; and (2) the patient underwent visual acuity testing, refractive examination, slit-lamp examination, and prism alternate cover testing. BPPV, was defined using the ICD-10 code H811. A diagnosis of BPPV was confirmed only if the patient also underwent a vestibular function test and was treated with canalith repositioning therapy.

Exclusion criteria were as follows: (1) participants with records of both amblyopia and BPPV in whom BPPV was diagnosed prior to or simultaneously with amblyopia; (2) participants first diagnosed with amblyopia during the follow-up period (2005–2013), to ensure temporal clarity between exposure and outcome. The starting point for the amblyopia group was the date of amblyopia diagnosis, and the end point was set as the date of first diagnosis of BPPV if a record of BPPV diagnosis existed or December 31, 2013 if a record for BPPV diagnosis did not exist.

Propensity score matching and selection of the control (non-amblyopia) group

The control group was assembled by 1:1 Propensity Score (PS) matching considering six independent variables (age, sex, residential area, economic status, history of hypertension (HTN) and history of diabetes mellitus (DM). PS matching was performed using a ‘greedy nearest neighbor’ algorithm with a 1:1 ratio. Because the baseline distributions of these variables were similar prior to matching, 1:1 matching was chosen to minimize selection bias while maintaining comparability between groups. Whether PS matching was successful or not was confirmed by the absence or presence of major imbalances in the standardized mean differences (SMDs) or figure in each group. After matching, all SMDS were below 0.1, confirming adequate balance.

Details of the patients’ age, sex, residence, household income, and comorbidities were obtained from the database. The study population was divided into 4 age groups (≤ 19, 20–39, 40–59, and ≥ 60 years). Household income was also assessed in 2 groups (low, 0%-69% of the median; high, 70%-100% of the median). For analyses, the residential area was divided into 3 groups based on home residence population density (Seoul, the largest metropolitan region in South Korea; other metropolitan cities in South Korea; and small cities and rural areas). We analyzed comorbidities that could be risk factors for BPPV: hypertension (ICD-10 codes I10 and I15) and diabetes mellitus (ICD-10 codes E10, E11, E12, E13, E14, O24). We defined the presence of comorbidities as any diagnoses of these codes between January 1, 2002, and December 31, 2004, prior to the diagnosis of amblyopia. The starting point for individuals in the control group was the date of their first visit to any hospital between 2002 and 2004 and the end point was decided as the date of first diagnosis of BPPV if there was a record of BPPV diagnosis or December 31, 2013 if there was no record of BPPV diagnosis.

Outcome variables and statistical analysis

After performing 1:1 PS matching, we conducted Cox proportional hazards regression analysis to estimate hazard ratios (HRs) for the development of BPPV. The univariate model assessed the crude association between amblyopia and BPPV using a single independent variable. The unadjusted multivariable model accounted for the matched data structure, and robust standard errors (sandwich estimator) were applied to prevent underestimation of variance resulting from the matched design. The adjusted multivariable model further included covariates used in the PS matching – age, sex, residential area, economic status, HTN, and DM – to obtain adjusted HRs.

Incidence rates were calculated as the number of new BPPV cases divided by the total person-years of follow-up and were expressed per 10,000 person-years. The cumulative hazard ratio was obtained through Kaplan–Meier survival analysis, and the R 3.5.3 statistical program (R Foundation for Statistical Computing, Vienna, Austria) was used to analyze the results. The SAS program (SAS Institute Inc, Cary, NC, USA) was also used for analysis of the NHIS-NSC database. All statistical analyses were performed between October and December 2024.

Effect estimates by demographic and clinical variables

We analyzed clinical variables to assess how the following factors are associated with the development of BPPV in the amblyopia group and the non-amblyopia group selected through PS matching. We also analyzed the association of sex, economic status, age, residential area, and comorbidities (including hypertension and diabetes mellitus) with the development of BPPV using a Cox regression model.

Ethical considerations

This study was approved by the Institutional Review Board of Jeonbuk National University Hospital (IRB No. 2022-02-023). The researchers obtained ethical approval to conduct this study in accordance with the Declaration of Helsinki. The requirement for informed consent was waived by the Institutional Review Board of Jeonbuk National University Hospital.

Results

We analyzed 1 million cohort samples and included a total of 5,320 individuals (2,660 in the amblyopia group and 2,660 in the non-amblyopia group), who were followed for a total of 9 years from January 2005 to December 2013.

Validation of PS matching

Table 1 presents the results of propensity score matching between the amblyopia (experimental) and non-amblyopia (control) groups based on six independent variables. All standardized mean differences (SMDs) were close to zero, indicating that the matching procedure achieved excellent balance between the two groups. As shown in Fig. 1, the distributions of propensity scores were nearly identical, confirming that the PSM was successfully performed.

Table 1 Demographics and clinical characteristics of amblyopia and non-amblyopia groups.
Full size table

Incidence rate and HR for BPPV

Table 2 shows the relationship between the 6 independent variables for BPPV and three statistics: 10,000-person years (incidence rate), unadjusted HR, and adjusted HR. The incidence rate (incidence per 10,000-person years) is a measure of the frequency with which a disease occurs over a specified time period. During the follow-up period, a total of 102 incident cases of BPPV were identified: 67 cases in the amblyopia group and 35 cases in the non-amblyopia group. The total follow-up duration was 47,725 person-years. As shown in Fig. 2; Table 2, the HR for BPPV in the amblyopia group was 2.25 (95% confidence interval (CI): 1.49–3.41), indicating that BPPV was 2.25 times more common in the amblyopia group than in the control group.

Table 2 Incidence rate and hazard ratios for the development of BPPV in the amblyopia and non-amblyopia groups.
Full size table

Effect estimates by demographic and clinical variables

In the clinical variables for demographic factors, the HR was 0.47 (95% CI: 0.31–0.73) in the subgroup of male patients compared with female patients. We also found that increasing age was significantly associated with the prospective development of BPPV (age 20–39: HR, 3.35; 95% CI: 1.93–5.83; 40–59: HR, 9.92; 95% CI: 6.05–16.28; and age ≥ 60 years: HR, 14.8; 95% CI: 7.65–28.69) (Table 2; Fig. 3). However, we observed no statistical differences in the risk of developing BPPV depending on the region of residence, economic status and comorbidities. Specifically, the HR for urban vs. rural residence was 1.12 (95% CI: 0.76–1.66), for low vs. high economic status was 0.95 (95% CI: 0.65–1.39), for hypertension was 1.09 (95% CI: 0.73–1.63), and for diabetes mellitus was 0.88 (95% CI: 0.58–1.35), indicating no statistically significant associations.

Fig. 1
figure 1

Distribution of propensity scores in the amblyopia (AM) and non-amblyopia (Non-AM) groups after matching, demonstrating adequate covariate balance. All standardized mean differences were close to zero, indicating successful matching. The cohort included 2,660 individuals in each group, and participants were followed for a 9-year period from January 2005 to December 2013. AM, amblyopia group; Non-AM, non-amblyopia group; HTN, hypertension; DM, diabetes mellitus.

Full size image
Fig. 2
figure 2

Overall cumulative hazard ratio (HR) for benign paroxysmal positional vertigo (BPPV) in the amblyopia and control groups over a 9-year follow-up. The adjusted hazard ratio for benign paroxysmal positional vertigo in the amblyopia group was 2.25 (95% confidence interval: 1.49–3.41), indicating a significantly higher risk compared to the control group. The cohort consisted of 2,660 individuals in each group, followed from January 2005 to December 2013. HR, hazard ratio; BPPV, benign paroxysmal positional vertigo.

Full size image
Fig. 3
figure 3

Forest plot presenting adjusted hazard ratios and 95% confidence intervals for benign paroxysmal positional vertigo in predefined subgroups. A higher risk was observed with increasing age and among women, while no significant differences were noted by region, economic status, or comorbidities. The analysis included 2,660 individuals per group, followed over a 9-year period from January 2005 to December 2013. CI, confidence interval; HTN, hypertension; DM, diabetes mellitus.

Full size image

Discussion

The results of this study provide substantial evidence of a significant association between amblyopia and an increased risk of BPPV. With an adjusted HR of 2.25, our data indicate that individuals with amblyopia face more than double the risk of BPPV compared to those without the condition. This link was particularly strong among older adults and women, suggesting that age-related or hormonal factors may further amplify this risk.

These findings suggest that amblyopia affects more than just visual acuity; rather, it reflects broader neurological dysfunctions, potentially disrupting the processing of sensory information in the central nervous system5,13. Across 47,725 person-years, we observed 102 incident BPPV cases, with amblyopia conferring a 2.25-fold increased risk (95% CI: 1.49–3.41), independent of metabolic, socioeconomic, and regional factors. This association was particularly pronounced among adults aged ≥ 60 years (HR 14.80) and in women, suggesting additive effects of age-related and hormonal factors.

These epidemiologic findings align with mechanistic evidence. Abnormal visual input during the critical period of development in amblyopia leads to persistent deficits in neural plasticity, affecting brain regions involved in sensorimotor coordination13,14. Such multisensory integration—especially between visual and vestibular inputs—is essential for maintaining balance and spatial orientation. Although amblyopia is a cortical visual disorder and BPPV is a peripheral vestibular condition, impaired visual feedback may disrupt central compensatory mechanisms in response to vestibular disturbances12,15,16. Binocular vision abnormalities, including reduced stereopsis, have been associated with increased postural sway even in the absence of vestibular dysfunction17,18. Neuroimaging studies further reveal altered cortical connectivity beyond the visual cortex in amblyopia6,7,19. Such impairments in multisensory integration may hinder central adaptation processes involving otolith signaling, potentially increasing susceptibility to BPPV. While our findings demonstrate a statistically significant association between amblyopia and BPPV, this should not be interpreted as evidence of a direct causal relationship. Given the mechanical nature of BPPV, amblyopia is unlikely to increase the occurrence of BPPV itself. Instead, we propose that individuals with amblyopia may exhibit less efficient central compensation or prolonged vestibular symptoms when BPPV occurs, due to impaired visual feedback and altered multisensory processing. This interpretation aligns with previous studies demonstrating reduced postural stability and disrupted visuo-vestibular integration in amblyopia. Accordingly, our findings should be regarded as hypothesis-generating, reflecting potential multisensory vulnerabilities rather than a direct etiologic link between the two conditions.

Furthermore, amblyopia has been associated with broader cognitive and motor deficits, including difficulties in spatial awareness, coordination, and motor control. These deficits could make it harder for those with amblyopia to adapt quickly to sudden balance disruptions, increasing their susceptibility to BPPV20,21. This cross-system interaction reinforces the importance of developing care strategies that integrate both visual and vestibular health for individuals with amblyopia.

The heightened prevalence of BPPV in older adults with amblyopia highlights a compounded risk, likely due to age-related decline in both visual and vestibular functions. Vestibular function naturally deteriorates with age, and when coupled with the sensory integration deficits seen in amblyopia, this decline may render older individuals particularly vulnerable to vertigo and balance disorders22,23. For clinicians, this finding suggests that older amblyopic patients may benefit from proactive vestibular assessments and balance-related interventions to mitigate fall risk and improve quality of life.

The increased risk of BPPV in women observed in our study is consistent with previous epidemiological data indicating higher BPPV prevalence in women across all age groups10. This supports the rationale for targeted screening and prevention strategies in high-risk populations such as older women. Sex-based difference may be partly explained by hormonal factors. Estrogen receptors have been identified in the inner ear, and estrogen fluctuations may affect calcium metabolism and otoconia stability24,25. Additionally, differences in vestibulo-ocular reflex gain, otolith mass, and vestibular-evoked myogenic potentials between sexes suggest that intrinsic physiological or mechanical variations may contribute to this disparity26,27. These mechanisms may help explain the greater vulnerability of women with amblyopia to BPPV.

Despite the strengths of this study, including its large sample size and rigorous propensity score matching, several limitations must be acknowledged. The use of diagnostic codes to identify amblyopia and BPPV introduces the potential for misclassification bias. Additionally, as a retrospective study, our findings suggest but do not confirm causality. Although individuals with known vestibular disorders were excluded at baseline, the possibility of subclinical vestibular dysfunction—such as vestibular-induced nystagmus—contributing to abnormal visual development in early life cannot be entirely ruled out. While the HR indicated a statistically significant association between amblyopia and BPPV, the absolute incidence of BPPV in both groups was relatively low, suggesting that the clinical impact may be limited. Furthermore, the NHIS-NSC database does not provide information on the severity of amblyopia, limiting stratified analysis based on visual acuity. It also lacks clinical details on several known risk factors for BPPV, such as head trauma, ear surgeries, and prolonged immobilization, thereby limiting our capacity to adjust for these potential confounders. As with other large-scale claims-based cohort studies, the inability to capture all relevant clinical factors represents an intrinsic limitation of this research design. Nevertheless, we believe that applying rigorous propensity-score matching and multivariable modeling in a nationwide cohort provides a statistically robust approach for identifying meaningful associations and generating hypotheses for future clinical studies. Therefore, our results should be interpreted as demonstrating a statistical association rather than a causal relationship between amblyopia and BPPV.

Surveillance bias is also possible, as amblyopic patients might undergo more frequent ophthalmologic evaluations, increasing the likelihood of BPPV diagnosis. In addition, subgroup analyses conducted across multiple strata may increase the risk of type I error due to multiplicity. As interaction terms were not formally tested in this study, the subgroup findings should be regarded as exploratory and interpreted with caution. These findings should therefore be interpreted with caution. Future prospective studies incorporating functional MRI and vestibular testing may help clarify the multisensory pathways linking amblyopia and BPPV.

In conclusion, this study highlights the significant association between amblyopia and an increased risk of BPPV, particularly in older adults and women. These findings suggest that visual and vestibular functions may be more closely linked than previously recognized, underscoring the importance of a multidisciplinary perspective that considers the broader neurological impact of sensory disorders such as amblyopia on overall health. However, given the small absolute number of BPPV cases, the observational nature of the study, and the potential for unmeasured confounding, these findings should be interpreted with caution. The results are best viewed as epidemiological and hypothesis-generating, providing population-level insight into possible visual–vestibular interactions rather than evidence for direct clinical recommendations. Future prospective studies are warranted to clarify the neural and sensory pathways linking amblyopia and vestibular dysfunction.