Dizziness and vertigo are common symptoms in Corona Virus Disease-19 (COVID-19) infections and have been widely described, especially in infections with wildtype Severe Acute Respiratory Syndrome Corona Virus-2 (SARS-CoV-2) virus (1, 2). Various pathomechanisms have been proposed as potential underlying causes for this symptom, ranging from direct neuroinvasion or neuropathy to hypoxia or hypercoagulopathy (3, 4). Among these reasons, autonomic dysfunction has been frequently observed in patients during and after COVID-19 infection (5–7). Additionally, otologic symptoms have been reported to be associated with COVID-19 (8). For example, a recent study by Tan et al. showed worse high-frequency audiometry and reduced otolith and semicircular function as measured by video head impulse test (vHIT) and vestibular-evoked myogenic potential (VEMP) testing (9). However, Tan's study included only relatively young individuals with COVID-19 (mean age, 28.98 years) and a rather small cohort of 26 individuals; the results, therefore, might not be universally transferable. In addition to this study, cases of vestibular neuritis and sudden-onset sensorineural hearing loss after COVID-19 have been described (10–12).

Therefore, it still seems to remain unclear which effect the SARS-CoV-2 virus might have on the vestibular system, both in the context of acute infection as well as in the context of long-term effects. Furthermore, the location of any possible damage that SARS-CoV-2 may cause to the vestibular system needs to be further examined, as the peripheral vestibular end organs or central vestibular structures might both be considered as sites of predilection. The aim of the present study, therefore, was to objectively measure vestibular function in patients with persistent complaints of dizziness after COVID-19 infection. Additionally, we carried out a retrospective review of patients who presented with acute vestibular syndrome (AVS) to our emergency room between September 2020 and March 2022 and who proved to have a positive polymerase chain reaction COVID-19 test with no further symptoms of disease other than vertigo.

In our study, which was conducted at a tertiary referral center, we included 50 patients aged 18 years or older with COVID-19 wild-type infection from May to November 2020. COVID-19 diagnosis was confirmed by polymerase chain reaction testing from swab samples. Patients with a previous history of vestibular disorders were excluded from study participation. The study design was an observational, cohort study design based on a convenience sampling method.

For evaluation of vestibular symptoms, subjects filled out the Dizziness Handicap Inventory (DHI) (13) during COVID-19 infection and then at two points of follow-up: at 6 months and again at 9–12 months after COVID-19 infection. Additionally, horizontal vHIT and subjective visual vertical (SVV) testing were performed in all participants 6 months after infection. If SVV testing showed abnormal results, VEMP testing was performed. Moreover, pure-tone audiometry was performed in all patients to rule out a significant hearing loss. If any of the included subjects presented with abnormal laboratory vestibular testing results at the 6-month post-infection visit, vHIT, SVV, and VEMP testing were repeated at the 9- to 12-month follow-up. In addition, vestibular testing results of COVID patients were compared with pre-existing normative data of healthy controls.

Moreover, a retrospective data analysis was performed in patients who presented with AVS at the emergency room between September 2020 and March 2022. Further descriptive patient characteristics were collected on all patients who could be identified who presented with AVS, had no COVID-specific symptoms, but simultaneously had acute COVID-19 infection diagnosed by routinely performed swab samples. The study was approved by the local ethics committee (#2020-549N) and conducted according to the Declaration of Helsinki; all patients gave written informed consent to participate.

Fifty patients [23 men (46%) and 27 women (54%)] were enrolled in the study, with an average age of 45 years (±15.44 years). In the course of COVID-19, most patients were treated as outpatients [44 patients (88%)], whereas only 5 patients (10%) were admitted as inpatients, including 2 patients (4%) in the intensive care unit (1 patient did not provide any information about his disease course). All 44 patients who received outpatient treatment considered their infection as “mild”, and 36 patients (72%) reported olfactory and gustatory symptoms. The other patient who received inpatient or ICU treatment described their infection as “severe”. All of them were admitted to hospital due to pneumonia.

As for medication, 11 patients (22%) reported to take hypertension medications, 4 patients (8%) antilipemic medication, 5 patients (10%) antidepressants, and 3 patients (6%) received asthma medications and analgesics.

Vestibular symptoms were mentioned by 23 (46%) patients and started at a mean of 2.96 days after COVID-19 diagnosis. Dizziness was categorized as non-specific by all affected, and 14 participants (28%) described worsening of dizziness when standing. Rotational vertigo was not reported by any of the participants. Three patients had to be excluded, as they had been suffering recurrent spells of dizziness before their COVID-19 diagnosis, bringing the total number of patients to 47 (Table 1).

Subjective dizziness was assessed in all 47 patients. The DHI score at diagnosis was 16.04, and 6 months after infection it significantly declined to 6.16 (paired one-side t-test, p < 0.000). The DHI score still significantly declined after 9–12 months post-COVID-19 infection (paired one-side t-test p = 0.013) compared with DHI scores during COVID-19. Nevertheless, after 9–12 months, DHI scores slightly (but not significantly; t-test p = 0.053) increased to 7.87 compared with the 6-month DHI score. Overall, women were more likely to experience dizziness than men (t-test, p < 0.05) and their symptoms tended to be more severe (DHI score 19.9, compared with 11.1 in men) and were more likely to persist 9–12 months after infection (Table 2).

Overall median horizontal gain was 1.01 for the right semicircular canal and 1.01 for the left (n = 47, female = 24, male = 23). The SVV testing showed a median deviation from true vertical of 0.9° to the right and 0.3° to the left. All vestibular testing results of COVID patients and pre-existing normative data of healthy controls are shown in Table 3, Figure 1. Only one female patient showed pathological vHIT (bilateral vestibulopathy; horizontal gain reduced to 0.45 right and 0.43 left with clear catch-up overt saccades in both directions) and SVV (4.5° and 5.1° deviation) results. In this patient, cVEMP and oVEMP were performed, but in both tests, no reproducible responses could be measured and were therefore interpreted as “absent”. Interestingly, this patient had a protracted course of infection, requiring intensive care, and also showed bilateral recurrent nerve palsy since the COVID-19 infection. In this patient, 6-month follow-up showed no significant improvement either in vHIT (horizontal gain 0.49 to the right side and 0.43 to the left side, persistent overt saccades) or in VEMP testing. Overall, there was no statistical difference in SVV, VEMP, or vHIT results between genders. COVID patients did not show declined semicircular or otolith function compared to pre-existing data of healthy controls (Table 3, Figure 1). Interestingly COVID-19 patients even showed better vestibular testing results (vHIT right and left: p < 0.0001; SVV right: p-value 0.0014 and left p-value < 0.0001).

Dizziness is a common symptom in patients with COVID-19 infection (2). Numerous reports describe audio-vestibular symptoms associated with COVID-19 (17–19). Nevertheless, even after more than 2 years of the pandemic, the etiopathogenesis of the underlying cause of dizziness remains unclear. It is conceivable and hypothesized that there is a direct effect on the cochlear and vestibular peripheral end organs and/or a neural irritation in the sense of a neuropathy or a virus-mediated immune response. The former is favored by the fact that SARS-CoV-2 viral proteins were detected in cranial nerves originating from the lower brainstem and in brainstem cells (20). Blood vessels, nerves, and the meninges also have been proposed as possible entry routes for the virus (21, 22). To better understand the manifestations of dizziness associated with COVID-19 infection, we combined a retrospective data analysis with a prospective study design.

In our prospective study, we could confirm that patients suffer from dizziness during infection, as DHI score was significantly elevated compared with our two follow-ups, 6 months later and then 9 to 12 months later. Moreover, almost 50% of the participants complained about dizziness, which occurred in a median of 2.96 days after infection. This result is comparable with other studies published recently (23). Women were significantly more often affected than men with dizziness during and after infection, and their DHI scores were significantly higher. This finding does not seem surprising, as women in general are affected more often by vestibular disorders and dizziness-related reduced of quality of life (24, 25).

In contrast to the subjective DHI and self-assessment, vestibular testing did not reveal any significant hypofunction in our study population. Hence, SARS-CoV-2 does not appear to cause long-term hypofunction or dysfunction of the vestibular system, as normal vestibular function was measured by vHIT and SVV testing. COVID patients did not show reduced vestibular function compared to healthy controls. Normal gain values were assessed, and no pathological saccades were seen in all but one patient. Interestingly, the existing literature is inconsistent, which is confirmed especially in two recent meta-analyses (22, 26). The study situation is difficult to compare, as most studies only collect subjective or objective parameters and differ greatly in their methodology. A recent study by Tan et al. describes significantly lower gains compared with healthy controls. However, it must be emphasized that the gains of the COVID group are still in the normal range, and it is questionable whether a small cohort of 26 individuals has sufficient sample size to generalize the findings (9). The same criticism applies to the study by Demir et al. that examines audio-vestibular function in a pediatric COVID collective comprising 35 children age 9–15 years (27). In our cohort, we did not detect overall reduced semicircular canal function or otolith dysfunction. Only one patient presented with bilateral vestibulopathy. Interestingly, this patient had a protracted course of infection, requiring intensive care, and showed bilateral recurrent nerve palsy since the COVID-19 infection. The patient's follow-ups did not show any significant improvement neither in terms of her vestibular system, nor the vocal cords. Usually, in patients with cranial nerve involvement, COVID-19 infections seem to be mild (28), which was not the case for the patient included in this study. However, it remains unclear, whether this is a viral- or drug-induced cranial nerve neuropathy, as the patient underwent intensive care and antiviral therapy.

In the retrospective part of our study, we identified 6 patients with acute peripheral vestibulopathy and PCR testing. Three of them developed COVID-typical symptoms, whereas three of them remained asymptomatic. Uni- or bilateral peripheral vestibulopathy have been described after COVID-19 infection as well as after vaccination (12, 29, 30). Nevertheless, in a recent prospective study by Gerb et al. an increase of central or peripheral vestibular syndromes after COVID-19 vaccination was not observed (31) and Charpiot et al. also did not find a correlation between AVS and COVID-19 (32). In conclusion, it can only be assumed but not proven that SARS-CoV-2 might cause acute vestibulopathy. In our case series, we also identified two patients with posterior circulation stroke a number which seems to be rather large in this small cohort consisting of eight patients. The predominant cause of PICA infarcts generally is an arterial-arterial thrombo-embolism from extracranial large arteries followed by cardio-embolism or in-situ disease (33). The relatively large number of PICA infarctions might be attributable to the systemic effects of COVID-19, particularly possible COVID-related hypercoagulopathy (34, 35). The triggering of coagulopathy in COVID-19 seems to be linked to alveolar damage resulting in an inflammatory storm with activated production of inflammatory cytokines including IL-6, resulting in generation of pro-coagulative factors and damage to endothelium as a prerequisite for thrombosis or arterial to arterial thrombo-embolism (36, 37).

A predilection for posterior infarcts in COVID-19 has also been described by Ahmed et al. (38). The authors hypothesize, that also the anatomical vascular peculiarity and variations can pose a more viable risk to isolated PICA lesions (38).

Nevertheless, due to the pandemic status of COVID-19, AVS and COVID-19 can also be merely a coincidence.

Overall, a structural affection of the vestibular system by SARS-CoV-2 seems to be unlikely and could not be confirmed in our study. Hence, there must be other possible explanations why individuals with COVID-19 experience dizziness: hyperventilation due to (subjective) dyspnea for example, can lead to dizziness and light-headedness which is often seen in patients with anxiety disorders (39). Cardiovascular or orthostatic factors, especially in combination with fever and a potential hypovolemia might also lead to dizziness. At the same time, autonomic dysfunction (i.e., orthostatic hypotension) is also a frequently observed symptom in COVID-19 patients (5–7) and has to be taken into consideration, especially as the dizziness was categorized as non-specific and worse in standing position. Moreover, a somatoform, stress-related component needs to be considered. Also, or precisely because the origin of dizziness in patients with COVID-19 remains unclear, a structured otological and neurological assessment should take place.

Our study has limitations that need to be mentioned. With only 50 patients, our sample size is rather small, although statistically sufficient. Additionally, due to restrictions imposed by hospital hygiene, vestibular laboratory testing was not allowed during the actual COVID-19 infection, but only after the patients recovered; this is a common limitation that has also been described by Fancello et al. (22). However, there is a controversial discussion on how the DHI correlates with objective tests: While Yip et al. found no correlation between balance testing and DHI (40), there are several studies describing a moderate correlation (41, 42).

An even more differentiated objective vestibular assessment would have been desirable, including head tilt or off-axis SVV and o- and cVEMP testing in all patients and ideally baseline testing even prior to the infection. A longer follow-up period and routinely performed vestibular tests especially during infection and 6, 9-, and 12-months, respectively, would have also been desirable not only to detect even the subtlest vestibular hypofunctions, but also to identify patients with vestibular affection which may have already been recovered after 6 months.

Moreover, only patients with wildtype infection were included; therefore, our results only reflect the effect of wildtype COVID-19 infection and potentially might not be generalizable for all variants. To date, no study has investigated the impact of different SARS-CoV-2 variants on the sensory system. It would be interesting to investigate whether the different variants have different influences on the sensory system overall and on the vestibular system in particular.

Overall, a persisting structural affection of the vestibular system by SARS-CoV-2 seems to be unlikely and could not be confirmed by vHIT, SVV and VEMPS in our study. It seems possible but unlikely that SARS-CoV-2 induces acute vestibulopathy. Nevertheless, dizziness is a common symptom in patients with COVID-19, which should be taken and worked through seriously.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by Ethik-Kommission II der Universität Heidelberg Medizinische Fakultät Mannheim Theodor-Kutzer-Ufer 1-3 68167 Mannheim. The patients/participants provided their written informed consent to participate in this study.

LZ and SL recruited patients, performed statistical analysis, provided conceptual support, and wrote the manuscript. MB recruited patients and performed statistical analyses. BW and FJ recruited patients and helped interpreting the results. VH and SH supported writing and finalizing the manuscript from a neurological perspective. NR recruited patients and advised in writing the manuscript. AS provided conceptual support, supervised the study, and helped writing and finalizing the manuscript. All authors contributed to the article and approved the submitted version.