A multicenter retrospective cohort study on incidence and diagnostics in emergency department patients with acute vestibular syndrome

Article information

Clin Exp Emerg Med. 2025;12(2):148-155

Publication date (electronic) : 2024 July 19

doi : https://doi.org/10.15441/ceem.24.225

Renske Eveline Henriëtte Maria Bijl^,1, Domenique Wilhelmina Antonia Maria Zaunbrecher², Petra Mathilda de Muynck², Ryanne Eggink¹, Ronique Timmer¹, Evian Willems², Sam Koning², Marieke Saskia Sanders¹, Kim Ellis Jie²

¹Department of Emergency Medicine, St. Antonius Hospital, Nieuwegein, Netherlands

²Department of Emergency Medicine, Jeroen Bosch Hospital, ’s-Hertogenbosch, Netherlands

Correspondence to: Renske E. H. M. Bijl St. Department of Emergency Medicine, Antonius Hospital, Koekoekslaan 1, Nieuwegein 3435 CM, Netherlands Email: renske@moiety.me

Received 2024 March 29; Revised 2024 June 25; Accepted 2024 June 28.

Abstract

INTRODUCTION

Acute-onset continuous isolated vertigo or dizziness, defined as acute vestibular syndrome (AVS), is a relatively common main symptom observed in the emergency department (ED), present in 2.6 million ED visits (2.5%–3.3%) in the United States annually [1,2]. The incidence of dizziness is uncertain as it is described variably (e.g., vertigo, spinning sensation, lightheadedness). According to the GRACE-3 (Guideline for Reasonable and Appropriate Care in the Emergency Department, 3rd) guidelines, AVS is defined as acute dizziness with no other neurological symptoms or other cause of dizziness and that is continuously present and persists at the time of evaluation [3]. The incidence of AVS in ED patients in the Netherlands is unknown.

The etiology of AVS can be roughly categorized into peripheral or central. Peripheral causes, such as vestibular neuritis, are generally considered benign and constitute a majority of cases. In contrast, diagnosis of a central cause (mostly posterior circulation ischemic stroke) is less prevalent and may constitute a life-threatening neurologic event [4,5]. Studies report incidences of stroke as an underlying cause in 3% to 6% of cases [6–8], with two smaller studies reporting incidence ranges of 10% to 25% in selected populations of patients with at least one risk factor for cardiovascular disease (CVD) [9,10].

In clinical practice, especially in the ED, differentiating between peripheral and central causes is a difficult task. Posterior circulation strokes frequently present with “isolated” vertigo, i.e., without additional focal neurological signs, and lack well-established “red flags” that may indicate a central cause [11]. A standardized diagnostic guideline was not available prior to this study [12,13]. Various diagnostic tools, such as the HINTS (head impulse test, observation of nystagmus, test of skew) exam and several imaging modalities, have been evaluated. The HINTS exam, when performed by trained physicians, has been shown to be more accurate in identifying a posterior circulation stroke compared to magnetic resonance imaging (MRI) <48 hours after symptom onset [14]. However, the success of HINTS application by less-specialized physicians remains unclear [14]. Also, while computed tomography (CT) imaging is frequently performed for such patients, its sensitivity for early presenting acute ischemic stroke has been shown to be as low as 10% [15].

Patients presenting with AVS may be misdiagnosed with a peripheral cause instead of a central etiology, and vice versa [16]. As a result, stroke diagnoses may be missed, and this may impose a critical impact on morbidity and mortality, with indications for acute thrombolysis and secondary prevention (e.g., antiplatelet therapy) not identified [16]. In contrast, patients who are incorrectly diagnosed with stroke at initial presentation may be overtreated.

In the Netherlands, similar to other countries, consensus on a standardized diagnostic approach was lacking at the time of this study. To improve acute diagnostics and treatment of patients presenting with AVS, the first step is to provide insight into current performance. In addition, the incidence of AVS in ED patients in the Netherlands is unknown. This multicenter study aims to provide insight on AVS incidence and current diagnostic approaches among ED patients presenting with AVS in the Netherlands. Furthermore, we identify the extent to which diagnosis is revised after follow-up and its effects on overtreatment and undertreatment.

METHODS

RESULTS

Study population

A total of 149,661 ED visits of patients ≥18 years old was registered during the study period, whose 10,220 medical records matched the predefined search terms (Supplementary Table 1). Among them, 500 patients had AVS. The selection of the study population is shown in Fig. 1.

Fig. 1.

Flowchart of the study population selection. EMR, emergency medical record; ED, emergency department; AVS, acute vestibular syndrome. ^a)Or one of its related terms or synonyms (see Supplementary Table 1, which lists the search terms in Dutch and English).

Baseline characteristics

The baseline characteristics are shown in Table 1. Stroke patients were significantly older, more often male, had higher rates of hypertension, dyslipidemia, nicotine abuse, atrial fibrillation at ED presentation, and a history of CVD and atrial fibrillation were predominantly present in stroke patients.

Table 1.

Baseline characteristics (n=500)

Diagnosis at the ED and follow-up

Eighty-five patients (17.0%) were diagnosed with stroke at ED presentation, 285 (57.0%) were diagnosed as nonstroke cases, and diagnosis was unsure in 130 patients (26.0%) (Table 2). After follow-up, diagnosis was revised in 145 cases (29.0%). Final diagnosis after follow-up consisted of 89 stroke diagnoses (17.8%), 380 nonstroke diagnoses (76.0%), and 31 unsure diagnoses (6.1%). The stroke group included 86 ischemic strokes (96.6%) and 3 hemorrhagic strokes (3.4%).

Table 2.

Diagnosis at ED and follow-up (n=500)

Twenty-nine ED patients (5.8%) who were initially diagnosed as nonstroke (n=9) or unsure (n=20) were revised to a stroke diagnosis after follow-up. This group was defined as a group of “missed stroke” cases. In contrast, in 25 of 85 ED patients (29.4%) initially diagnosed with stroke, diagnosis was revised to nonstroke (n=20) or unsure (n=5) after follow-up.

Diagnostic approach

The diagnostic tools used at ED presentation and follow-up are shown in Table 3. At ED presentation, a HINTS exam was reported in 106 patients (21.2%). NCCT imaging was performed in 342 cases (68.4%), MRI was performed in 14 patients (2.8%), and CT angiography was performed in 9 patients (1.8%). During follow-up, MRI was conducted in 139 patients (27.8%), predominantly in stroke and unsure cases (P<0.001).

Table 3.

Diagnostic tools at ED and follow-up

Therapy and hospitalization

The therapies prescribed at ED presentation are shown in Table 4. Of 85 patients diagnosed with stroke at ED presentation, eight received thrombolysis. In three of these thrombolysis cases, the diagnosis was changed to nonstroke after follow-up. In addition, clopidogrel and acetylsalicylic acid were prescribed in 104 (20.8%) and 31 cases (6.2%), respectively (either diagnosed as stroke or unsure at ED presentation). In 72 of these patients (5.0%), the diagnosis was revised to nonstroke after follow-up, and this group inappropriately received antiplatelet therapy.

Table 4.

Therapy at ED presentation and hospitalization

Regarding therapy opportunities for 29 patients whose stroke diagnosis was initially missed at ED presentation, 10 would have been eligible for secondary prevention by antiplatelet therapy and 23 would have been eligible for thrombolysis (exclusively based on the evaluation of timeframe and contraindications) [17].

DISCUSSION

This multicenter study is the first to report on the incidence and diagnostic approach of AVS in ED patients in the Netherlands. The annual incidence of AVS is 0.1%. In addition, daily clinical practice showed underreporting of HINTS and overuse of NCCT. The final diagnosis was revised in nearly one-third of ED patients.

The annual incidence concluded in this study is considerably lower than that previously reported in studies from the United States (2.5%–3.3%) [1,2]. In the Netherlands, patients are generally seen first by a general practitioner, who then refers selected high-risk patients to an ED. Therefore, a substantial number of low-risk patients with AVS may be “filtered out” and will not present to the ED. In addition, our study used a strict symptom definition (e.g., only isolated vertiginous dizziness), whereas other studies used a broader definition of dizziness [18].

At ED presentation, 17.0% of patients with AVS were diagnosed with stroke. This is in line with previous studies that showed a stroke incidence ranging from 10% to 25% in selected populations of patients with at least one risk factor for CVD [9,10]. The “gatekeeper” function of Dutch general practitioners may explain this comparable high stroke rate in relatively higher risk patients in our study. Other studies reported remarkably lower incidences of stroke ranging from 3% to 6% [6–8], which may again be explained by the broader definition of dizziness that has been applied.

In this study, diagnosis at ED presentation was revised after follow-up in roughly one-third of cases. This is lower than the conversion rate of 44.0% reported by Royl et al. [5]. It should be noted, however, that follow-up data were only available in one-fourth of their patients. Due to our retrospective study design, our follow-up was dependent on the final reported diagnosis. It was assumed that, if patients did not return, the initial diagnosis was correct. In addition, this study showed a substantially lower incidence of missed stroke of 5.8% versus that of 35% reported by Kerber et al. [6]. In the study by Kerber et al. [6], patients were assessed by ED physicians, whereas, in this study, patients were either seen by an ED or neurology physician. It is unknown to what extent this may have influenced the conversion rate, but the lack of a true gold standard and patient selection may have contributed to this difference.

Regarding diagnostics, a HINTS exam was completed in only 21.2% of patients, while, based on the literature, this may now be considered the gold standard [3]. This is consistent with other reports (10%–30%) [12,19–21]. Due to our retrospective study design, we were dependent on reported HINTS exams, which might have resulted in an underestimation. Despite this, the noteworthy underuse may be due to neurologists and ED physicians not feeling sufficiently competent in applying or interpreting the HINTS exam. In addition, unfamiliarity with this examination may also play a role. Our findings underline the importance of better education in performing HINTS exams in the ED.

NCCT imaging was conducted in 68.4% of ED patients. This rate is considerably higher than that reported by Saber Tehrani et al. [22], who found that 39.2% of ED patients with dizziness underwent NCCT imaging in the United States. This may be due to the different selection of high-risk patients in the Dutch setting compared to the United States. However, this imaging modality shows poor performance in detecting posterior circulation strokes, with a sensitivity of just 7% to 10%, and may lead to physicians being falsely reassured of a peripheral etiology [15]. Presumably, the main reason for conducting NCCT imaging is to rule out intracranial hemorrhage. However, in ED patients with dizziness without focal neurological symptoms, screening for intracranial hemorrhage rarely shows a positive result. The extensive use may be due to its ready availability and lower cost compared to MRI.

One-third of patients underwent MRI, with the majority of imaging performed during follow-up versus only 2% in the ED. The relatively low number of MRI scans performed in the ED is in accordance with the findings by Saber Tehrani et al. [22] and may primarily be attributable to the organization of patient flow from the ED to admission in the Netherlands. In Dutch EDs, MRI availability is relatively limited, and MRIs are typically conducted after patient admission to the ward rather than during the initial ED visit to minimize patient stay times at the ED. Additionally, although delayed MRI has greater sensitivity in detecting cerebral infarctions compared to early MRI [15], it remains to be determined whether increasing the number of MRIs is necessary if the HINTS protocol is more effectively taught and implemented.

A stroke was missed at ED presentation in 5.8% of cases. One-third of these cases would have been eligible for additional antiplatelet therapy. Furthermore, exclusively based on the evaluation of timeframe and contraindications, 79.3% of missed stroke patients would have been eligible for thrombolysis [17]. Although untreated ischemic stroke is associated with increased morbidity and mortality, performance of thrombolysis in AVS must be determined. In the case of mild symptoms, administration of thrombolytics may not be justified due to its potential severe side effects [23].

More than 50% of ED patients were admitted for inpatient observation. A majority of these patients (79.2%) had an unsure diagnosis at ED presentation, whereas a minority had underlying stroke. Earlier differentiation of AVS etiology may prevent costly admittance to the hospital.

Recently, a guideline focusing on this particular patient group was published [3]. Our study substantiates the need for application of these guidelines, which may help to improve knowledge and training of emergency physicians on this subject. For instance, training of emergency physicians in performing the HINTS exam could minimize unnecessary use of CT or MRI. These adaptations may also lead to an important cost reduction and, hopefully, a better selection of patients suffering from a stroke who require treatment and adequate follow-up.

An important limitation of this study includes its retrospective design, which allows missing data and selection and reporting biases due to non-blinding of researchers for outcome. The lack of a true gold standard for the diagnosis is a challenge. We consider the HINTS examination, when performed by adequately trained physicians, to be the most appropriate test (GRACE-3). However, in our study, only 21% of patients underwent the HINTS examination. We had to comply with the best available diagnosis. In addition, data on follow-up were only available from the two teaching hospitals. Although unlikely, if patients presented at another medical care facility, it is unknown whether alternative diagnoses were made, possibly implying an underestimation of missed strokes.

Another limitation is that we focused exclusively on AVS patients, i.e., patients with continuous dizziness. Patients with episodic (non-continuous) dizziness/vertigo may also have had a transient ischemic attack. Finally, it remains a challenge to conduct research on dizziness, given the highly variable descriptions of symptoms and the evolving definitions in recent years. We used strict inclusion criteria, which may have led to an underestimation of incidence.

Despite the limitations, an important strength of this study is its multicenter design. To our knowledge, this is the first multicenter study to provide insights into the incidence and diagnostic pathway of AVS within Dutch EDs. Additional strengths encompass the substantial study cohort and comprehensive assessment of eligibility by multiple evaluators.

In summary, the annual incidence of AVS in Dutch EDs is 0.1%, with a central etiology in roughly one-sixth of ED patients. The initial diagnosis was revised after follow-up in nearly one-third of cases. This study substantiates the clinical difficulty of differentiating stroke from nonstroke causes in the ED. Daily clinical practice lacks a uniform diagnostic approach, with only a minority of patients undergoing a HINTS exam and overuse of NCCT. With the current diagnostic workup, strokes are missed, and patients are at risk for overtreatment and undertreatment. To improve patient care, further research on an optimal and uniform diagnostic approach is warranted.

Supplementary materials

Article information Continued

Notes