Course of Duration and Trigger Factors of Vertigo Attacks in Patients with Benign Recurrent Vertigo, Menière’s Disease, or Vestibular Migraine Open Access

Benign recurrent vertigo (BRV), Menière’s disease (MD), and vestibular migraine (VM) show many similarities with regard to the course of vertigo attacks and clinical features, as was clearly illustrated in two previous studies performed in our center [van Esch et al., 2017; van Leeuwen et al., 2022]. A favorable course after 3 years of follow-up applied to the majority of the patients in the three diagnosis groups, with 67–71% of the patients being free of vertigo attacks [van Leeuwen et al., 2022]. The above syndromes can be roughly distinguished by isolated vertigo attacks (BRV), vertigo in combination with ear involvement (MD), and vertigo in combination with headache (VM).

From the physician’s perspective, there are various reasons to assign a correct diagnosis to a patient with vertigo attacks – primarily to ascertain the cause of the presented complaints, to inform the patient properly, to predict the course of disease, and also to establish appropriate treatment. In recent years, good progress has been made in acquiring knowledge about diagnosis and treatment for MD, VM, and BRV. Although some research on the course of disease in terms of vertigo attacks has been conducted in MD [Perez-Garrigues et al., 2008], due to the lack of evidence-based data on the other two conditions, physicians are not always able to inform patients about their prognosis. Therefore, there is a need for empirical data on the duration and trigger factors of vertigo attacks in these conditions.

Research in patients with MD, VM, and BRV has established caffeine, alcohol, stress, sleep deprivation, dietary indiscretion, hormonal change, allergies, head movements, physical activity, and barometric pressure change to be precipitating factors for vertigo attacks [Slater, 1979; Rauch, 2010; Lauritsen and Marmura, 2017]. If these factors are consistent triggers of attacks over time in individual patients, this offers opportunities to avoid these factors through providing advice or therapy.

Regarding the duration of vertigo attacks, reference is made to the literature on the clinical distinction based on duration, which is part of the diagnostic criteria for the above mentioned three conditions. In contrast to benign paroxysmal positional vertigo, the attacks in the above mentioned vestibular disorders last longer than seconds. Whereas vertigo attacks in MD generally last for 20 min to 12 h, VM is characterized by an attack duration ranging from 5 min to 72 h and BRV also has an attack duration ranging from 5 min to hours[Perez-Garrigues et al., 2008; Lempert et al., 2012; van Esch et al., 2017; van Leeuwen et al., 2022]. Taking into account the margin of vertigo attacks within the definition of the conditions, changes in duration of vertigo attacks must be defined differently for the three conditions. Nevertheless, if vertigo attacks cannot be completely avoided, a decrease in their duration is desirable for all three groups. This study aimed to explore to what extent the duration and trigger factors of vertigo attacks change during a 3-year follow-up period in patients with BRV, MD, or VM.

An existing dataset from a prospective observational study on recurrent vertigo was used. In this study, patients diagnosed with BRV, VM, or MD at the Apeldoorn Dizziness Center (ADC) in Gelre Hospitals location Apeldoorn between January 2015 and November 2016 were included [van Esch et al., 2017]. A total of 121 patients were included, consisting of 44 patients with BRV, 34 with (definite or probable) VM, and 43 with MD. When the study was designed (in 2014), differences in clinical presentation between BRV, VM, and MD were largely unknown. Thus, we did not a priori have a quantifiable hypothesis at the time and therefore decided to perform an explorative, hypothesis-generating study and a sample size calculation was not performed. Instead, we intended to include 40 patients for each of the three conditions studied.

Participants in this study had to meet the following criteria regarding one of the aforementioned diagnoses. As for definite and probable VM, the criteria according to Lempert et al. were used [Lempert et al., 2012]. The criteria as formulated by the Committee on hearing and equilibrium guidelines for the diagnosis and evaluation of therapy in MD (1995) had to be met as inclusion criteria for definite unilateral MD [Committee on Hearing and Equilibrium guidelines, 1995]. The revised diagnostic criteria for MD that were published in 2015 were not taken into account as the recruitment of participants had already started [Lopez-Escamez et al., 2015]. Since BRV is currently not formally recognized as a syndrome, the description for definite BRV by Slater had to be satisfied [Slater, 1979]. The aforementioned vestibular disorders were diagnosed by simultaneous assessment of patients by an otorhinolaryngologist and a neurologist at the ADC. For more detailed information about the inclusion and exclusion criteria, selection procedures, clinical examinations, and the conducted additional tests, we refer to two previous publications on the study [van Esch et al., 2017; van Leeuwen et al., 2022].

Prior to the follow-up period, measurements regarding baseline demographic and clinical characteristics were conducted. After a patient’s first visit to the ADC, he or she received a study-specific questionnaire by telephone interview at 3 months, 6 months, and every 6 months thereafter for a period of 3 years. All interviews were conducted by the same research nurse. During the follow-up surveys, the research nurse inquired in detail about the occurrence of vertigo attacks in the past 6 months, medication used for vertigo attacks (including intratympanic injections), additional symptoms during vertigo attacks, the frequency and duration of vertigo attacks, as well as their trigger factors.

If the patient had vertigo attacks in the past 6 months, their usual duration was expressed in seconds, minutes, hours, or days. To determine the difference in duration of vertigo attacks, the usual attack duration at the 12-, 24-, and 36-month follow-up measurements was compared to the baseline attack duration. At every follow-up measurement, the following trigger factors were inquired: stress (i.e., general life stress, not the fear of having an unpredictable attack), fatigue, menstrual cycle, nutrition, alcohol, head motions, physical activity, and sleep deprivation. In addition to the standard requested triggers, participants could also mention other trigger factors that were specific to their vertigo attacks. The spontaneously mentioned trigger factors were combined with the standard requested trigger factors (if applicable) through consensus between two of the authors (MCU and RBvL). For the change in trigger factors over time, we focused on the three most commonly reported triggers at baseline: stress, fatigue, and head movements [van Esch et al., 2017]. For each of these trigger factors, change was defined as the disappearance of the trigger at a biannual follow-up measurement when the patient had mentioned this particular trigger at baseline. In addition, we also examined whether the trigger factor concerned had been replaced by another trigger factor when the vertigo attacks still persisted at subsequent follow-up measurements.

Baseline differences between diagnosis groups were assessed by cross-tabulation and analyzed using χ² test or, in case of expected count of <5 in one or more cells of a contingency table, Fisher’s exact test. Between-group differences in age at the time of inclusion and age of onset of the first vertigo attack were tested using one-way ANOVA. Cochran’s Q test was used to analyze the proportions of patients with shortening in duration of vertigo attacks for the BRV, VM, and MD separately at the 12-, 24-, and 36-month follow-up measurements relative to baseline. For this purpose, the categories of “equal attack duration” and “longer attack duration” were combined before performing the tests.

The probability of a particular trigger factor to have disappeared at a subsequent follow-up visit was analyzed within each of the three diagnosis groups separately using generalized estimating equations models with a binomial distribution, logit link function, and an unstructured correlation matrix. Results are expressed as odds ratios (ORs) and their 95% confidence intervals per biannual measurement, with the baseline measurement serving as the reference (i.e., OR = 1). Age and sex were included as covariates to correct for differences in demographic characteristics that may have occurred due to varying composition of the diagnosis group at each time point because patients could either (re-)enter (in case of return of vertigo attacks) or (re-)exit (in case of disappearance of vertigo attacks) the analysis.

All analyses were performed using SPSS software (version 25, released 2017). A two-sided p value <0.05 was considered statistically significant. In addition to the statistical analyses, graphical plots were created to provide visual insight into the changes in the presence or disappearance of the most common triggers (i.e., stress, fatigue and head movements) over time in the individual patients in the three diagnosis groups.

The study population consisted of 121 patients divided over the three diagnosis groups (Table 1). The inclusion and follow-up of the study population are described elsewhere [van Leeuwen et al., 2022]. During follow-up, 4 patients (1 BRV, 1 MD, and 2 VM) discontinued their participation. The MD group was the only group that consisted of more males than females (p = 0.003). The patients in the BRV group were on average approximately 7 years older than the patients in the other two groups (p = 0.032). At baseline, betahistine was clearly more used in the MD group (58.1%) compared to the BRV (9.1%) and VM (8.8%) groups (p < 0.001). In addition to the medication mentioned at baseline, 23.3% of the MD patients received one or more intratympanic injections during their 3-year follow-up.

Table 1 also shows that at baseline all participants reported their vertigo attacks to last minutes to days, the largest proportions (i.e., BRV 45.5%, MD 81.4%, VM 44.1%) reporting the attacks to last for hours. A significant between-group difference in attack duration between the three diagnosis groups was observed (p = 0.003). Stress was the most common trigger factor in each group at baseline (BRV 40.9%, MD 62.8%, VM 76.5%; p = 0.005), followed by fatigue (BRV 31.8%, MD 48.8%, VM 70.6%; p = 0.003).

Table 2 shows the trend over time in terms of decreasing numbers of patients who reported to have vertigo attacks in each diagnosis group [van Leeuwen et al., 2022]. After 36 months, 30.2% of BRV, 35.7% of MD, and 37.5% of VM patients reported to have had attacks in the past 6 months. Among these patients, the proportion reporting shortened attack duration as compared to baseline did not significantly change at the consecutive annual follow-up measurements in any of the three diagnosis groups (Cochran’s Q test: BRV: p = 0.135, MD: p = 0.368, VM: p = 1.000).

Table 3 shows the ORs for stress being reported as trigger factor for vertigo attacks in the past 6 months at the consecutive biannual follow-up measurements relative to the baseline measurement per diagnosis group. The probability of VM patients reporting stress as a trigger of their attacks showed a consistent reduction up until the 24-month follow-up measurement, with ORs ranging from 0.15 to 0.33 (all p < 0.05). A similar pattern was seen for fatigue as trigger in the VM patients up until the 30-month follow-up measurement, with ORs ranging from 0.15 to 0.29 at the 6-, 12-, 24-, and 30-month measurements (all p < 0.05, see Table 4). In the MD group, a consistent reduction of head movements as trigger was observed from the 24-month measurement onward (ORs ranging from 0.07 to 0.11, all p < 0.05; see Table 5).

The composition of the diagnosis groups varied with regard to the presence or absence of vertigo attacks and their trigger factors during the consecutive measurements. Figure 1 shows the appearance and disappearance of stress as a trigger factor over time in individual patients in the three diagnosis groups. At baseline, 18 of the 42 BRV patients (42.9%) reported stress as trigger factor for their vertigo attacks, while 24 BRV patients (57.1%) mentioned stress as trigger at some point during follow-up (Fig. 1). In the MD patients, 26 out of 41 (63.4%) reported that stress triggered their vertigo attacks at baseline, whereas over the entire follow-up period, a total of 36 MD patients (87.8%) mentioned stress to be a trigger. For the VM patients, these numbers were 24 out of 32 (75.0%) and 28 (87.5%), respectively. Figure 1 illustrates that although many patients in the three diagnosis groups reported stress as a trigger factor at baseline, the pattern in which stress remained to be a relevant and consistent trigger over time varied strongly between patients. Online supplementary Figures 1 and 2 show the same plots for fatigue and head movements as trigger factors for vertigo attacks (for all online suppl. material, see https://doi.org/10.1159/000531545).

While our previous studies in this study cohort focused on the clinical characteristics of BRV, VM, and MD [van Esch et al., 2017; van Leeuwen et al., 2022] and the course of vertigo attacks in terms of their frequency [van Esch et al., 2017; van Leeuwen et al., 2022], the current analysis explored whether the duration and trigger factors of vertigo attacks in patients suffering from these conditions change over time. Our findings suggest that despite the decrease in the number of patients still experiencing vertigo attacks in all three diagnosis groups, no shortening of the duration of the attacks was observed in patients who continued to have attacks during the 3-year follow-up. In patients with MD and VM who reported to still have attacks during biannual measurements, a statistically significant decrease in the probability of stress, fatigue, and head movements playing a role as triggers of their attacks was seen.

Previous research performed by Perez-Garrigues et al. also focused on the duration of vertigo attacks in MD, and found a constant decrease in vertigo duration during a follow-up period of 30 years [Perez-Garrigues et al., 2008]. These authors differentiated between an attack duration of 20 min to 2 h, 2–6 h, and more than 6 h. They observed that the rate of vertigo attacks in the groups having attacks with a duration less than 6 h reduced progressively over time. The frequency of vertigo attacks lasting for more than 6 h did not change substantially. However, like in our study, the number of vertigo attacks of any length will decrease as the number of patients who keep having attacks decreases. For this reason we focused solely on patients who remained having vertigo attacks. In contrast to Perez-Garrigues, in their study, Havia and Kentala reported that the duration of vertigo attacks in MD patients from their tertiary dizziness clinic increased as time progressed [Havia and Kentala, 2004]. It should be mentioned, however, that Havia and Kentala used a cross-sectional instead of a longitudinal study design and categorized their patients based on disease duration. In our study we prospectively followed the patients for 3 years after their initial visit to our dizziness clinic, which enabled us to look at within-patient changes in attack duration over time. In addition, given the changing diagnostic criteria for MD in the past decades, the validity of the diagnostic criteria applied in their 2004 study makes their findings more difficult to interpret.

In an attempt to better understand the effect of health education on recurrence of VM, Xiao et al. found a 20% reduction in the duration of vertigo attacks in 48.5% of the VM patients in their study after lifestyle changes during a median follow-up period of 15 months [Xiao et al., 2019]. Mean attack duration decreased from 17.4 (SD 1.4) hours before to 10.5 (SD 0.9) hours after health education. However, it should be noted that the patients in their study lent themselves quite well to lifestyle modification, given the high percentage of patients with comorbid sleep disorders, lack of exercise, and stress at baseline [Xiao et al., 2019]. This implies that their study population may not be directly comparable to ours. To our knowledge, no previous studies have examined the change of vertigo attack duration in patients with BRV.

To our knowledge this is the first study investigating the change of trigger factors for vertigo attacks over time in the conditions studied. The modern conception of a MD-diseased ear as a fragile ear due to an impaired homeostasis, suggests some trigger factors that mostly are already targeted when the treatment starts. The theory states that hearing and balance functions become vulnerable to a complex of internal and external factors, such as stress, sleep deprivation, dietary indiscretion, hormonal change, allergies, and barometric pressure change [Rauch, 2010].

Stress has shown to be a trigger for almost all symptoms of MD, and the association with stress has also been reported for VM. In addition to stress, higher vertigo rates were also reported at lower mean air pressures in MD patients [Schmidt et al., 2017]. Like in VM, it is assumed that in BRV factors like alcohol, stress and sleep deprivation serve as triggers for vertigo attacks. Previous studies have looked at the temporal relationship between stress and VM, examining whether stress is the result of vertigo attacks or vice versa. On the patient population level attempts have been made to establish a temporal relationship between the occurrence of vertigo attacks and the stress experienced [Soderman et al., 2004; Kitahara et al., 2016; Basura et al., 2020]. Soderman et al. reported vertigo attacks to occur within 3 h of experiencing stress [Soderman et al., 2004]. However, no research has been reported on the course of stress and other trigger factors within patients over time [Neuhauser and Lempert, 2004]. Our current study shows that the most commonly reported factors (i.e., stress, fatigue and head movements) are not consistent over time in terms of triggering attacks in patients with MD and VM. For BRV we found no evidence to assume that these trigger factors will disappear in patients who remain to have attacks in the next 3 years. Brantberg and Baloh showed that vertigo attacks in patients with BRV cannot be distinguished from those in patients with MD on the basis of attack duration, triggers or associated symptoms other than auditory, but did not look at change of triggers over time [Brantberg and Baloh, 2011]. In order to get in-depth insight on the (dis)appearance of triggers in individual patients in the three diagnosis groups we added the visual displays in Figure 1 and online supplementary Figures 1 and 2. We can only speculate why some triggers (dis)appeared over time in some patients and not in others. Use of medication (see Table 1) and intratympanic injections during follow-up may have reduced their background risk of vertigo attacks and/or their sensitivity to certain triggers. Patients may also have become more proficient in avoiding their known triggers, for instance through psychological help for stress management. This makes changes at the individual level all the more interesting and challenging to study.

Clear strengths of our study are the well-defined diagnosis groups, the 3-year follow-up period and the large degree of completeness of the follow-up data. A limitation is the rather “crude” way of measuring attack duration and trigger factors using biannual telephone interviews. The choice not to ask patients to visit our clinic every 6 months for 3 years just for the study will have enhanced their willingness to participate and reduced the number of drop-outs and missing data, but at the same time did not allow us to perform additional on-site testing (e.g., pure-tone audiometry, extended Fletcher Index, videonystagmography) to monitor the course of disease as a part of the follow-up. We cannot tell whether or not patients with recurrent attacks were asymptomatic between attacks, as this was not specifically inquired during the follow-up telephone interviews. In hindsight, it would have been a valuable addition to the interviews to have asked this.

With regard to the duration of vertigo attacks, we categorized patients’ responses a priori in four categories (i.e., seconds, minutes, hours, or days). More precise estimates of the exact attack duration or trigger(s) using, for instance, a paper diary or a smartphone app [Martin et al., 2020] might have shown more subtle changes in duration or precipitating factors of vertigo attacks over time. On the other hand, one could wonder whether study participants would be prepared to record this information for 3 years in a row. Because patients were retroactively asked to specify the average duration and triggers of their attacks (if any), recall bias may have occurred.

Even though at the start the total study population was of reasonable size, due to the increasing number of patients who did no longer report vertigo attacks, the number of patients in the longitudinal analyses gradually reduced to smaller numbers. A final limitation is that the entire study population came from one tertiary center, which means that this may be a selection of patients with more severe complaints or complaints that are more difficult to interpret and manage. Despite these limitations, we feel that our study provides some valuable in-depth insights into the course of vertigo attacks over time that do need confirmation in future studies.

In conclusion, our explorative study showed no reduction in vertigo attack duration over time in patients with BRV, MD, or VM who remained to have vertigo attacks. Over time, stress, fatigue, and head movements became less predominant triggers for vertigo attacks in patients with VM or MD. Further research on change in vertigo attack duration in larger patient samples and with more specific measurement of duration and triggers of attacks in these diagnosis groups is needed.

The authors appreciate the time and effort the study participants dedicated to the study.

The study was designed and conducted in compliance with the Helsinki Declaration. The study protocol was reviewed and approved by Gelre hospitals’ Institutional Review Board, approval number TCO 2015 - 14.79. All patients gave written informed consent before entering the study.

At the time of submission the authors have no potential conflicts of interest to disclose.

No funding was received for this study.

R.B.v.L. and T.R.S. initiated the study. M.C.U. performed the statistical and graphical analyses and wrote the draft version of the paper. T.R.S. critically checked all statistical analyses and submitted the paper to the journal. C.C. collected the data and supported the analyses. R.B.v.L., M.C.U., C.C., and T.R.S. critically reviewed the draft version of the paper and agreed with the final version.

The data that support the findings of this study are not publicly available because the field labels and codebook are in the Dutch language, but are available from the corresponding author (T.R.S.) upon reasonable request.