Chronic vestibular dysfunction as an unappreciated cause of chronic nausea and vomiting

Abstract

In patients with chronic nausea and/or vomiting, gastroparesis is frequently diagnosed, often on the basis of abnormal gastric emptying scintigraphy (GES). When typical treatments fail, patients may be referred to a referral center. This retrospective study evaluated the diagnoses made in patients referred for chronic nausea and vomiting and appraised the GES utilized to assess these patients. Records of outpatients referred for chronic nausea and vomiting over a 3-year period were analyzed for previous evaluation and treatment, subsequent investigation, and response to treatment. Of 248 patients referred for chronic nausea and vomiting, 156 (63%) were referred with a suspected diagnosis of gastroparesis. Of 102 GES available for review, 95 were done with nonstandardized methods. Repeat standardized testing was normal in 27 of 36 patients (75%). Only 28 patients (11%) had confirmed gastroparesis. The most common specific diagnosis in the entire group was chronic vestibular dysfunction (CVD, 64 patients, 26%) made by abnormal modified Fukuda stepping test, nystagmus, or abnormal Romberg test. CVD patients did not typically report a history of an inner-ear disorder or vertigo. Eighty-nine percent of CVD patients were given trials of antivertiginous medications; of the 39 followed for a median of 5 months, improvement occurred in two thirds. Diagnosis of gastroparesis should not be based on a nonstandardized GES. In our referred patients, gastroparesis was infrequent, while CVD was much more likely. Treatment for CVD may mitigate the nausea and vomiting.

Chronic nausea and vomiting can be a challenging symptom complex for both the patient and physician for a variety of reasons, including the large number of potential underlying causes, the difficulty of making a specific diagnosis, and the difficulty of adequately treating symptoms. An accurate history and physical examination are paramount in formulating a diagnostic approach. Once structural lesions are excluded, an underlying motility disorder often is sought (1).

Gastric emptying scintigraphy (GES) is considered to be a critical test when gastroparesis is suspected. A 4-hour, solid-phase gastric emptying study using a standard protocol currently is recommended as the optimal test for evaluation of gastric emptying (2). Despite this recommendation, a variety of less-reliable alternative protocols are still used at some institutions. Emptying also may be altered by concomitant symptoms, such as nausea, or concurrent medications, such as narcotics. Unless the test is done properly, results may be misleading.

Even when the standardized exam is utilized properly, some patients with chronic nausea and vomiting have normal gastric emptying studies. This group of patients has recently been described as having chronic unexplained nausea and vomiting. These patients are clinically indistinguishable from those with delayed emptying in terms of symptom severity and health care utilization (3). The causes for symptoms in these patients are not known.

Patients with chronic nausea and vomiting often are referred to our institution for consideration of placement of gastric electrical stimulators. Many previously were diagnosed with gastroparesis, often on the basis of nonstandard gastric emptying scans. The aim of this retrospective study was to evaluate the diagnoses made in a consecutive series of patients referred for chronic nausea and vomiting and to appraise the GES utilized to assess these patients.

METHODS

Consecutive patients with chronic nausea and vomiting referred to a single physician at our institution from January 2008 to December 2010 were identified from billing records. Patients with symptoms for over a month were included. Exclusion criteria included age <18 years and pregnancy at the time of referral.

Medical records were reviewed to characterize the evaluation and treatment prior to referral, subsequent investigation and diagnosis, and therapeutic responses when available. This study was approved by the institutional review board in February 2011.

Clinical diagnoses were based on the following criteria:

• Gastroparesis was identified by chronic symptoms of nausea and/or vomiting, typically delayed after eating and accompanied by an abnormal standardized GES (>10% of test meal remaining at 4 hours).

• Suspected chronic vestibular dysfunction (CVD) was defined on the basis of symptoms of nausea and/or vomiting in conjunction with signs of altered vestibular dysfunction on physical examination. This included the presence of nystagmus, a positive Romberg test, or an abnormal modified Fukuda stepping test. We performed the modified Fukuda stepping test by having the patient march in place with eyes shut and ears occluded for 60 seconds. A turn of 90 degrees or more to the right or left was considered abnormal (see supplementary video online).

• Gastroesophageal reflux was defined by symptoms of heartburn and/or regurgitation in association with abnormal duration of reflux on pH-monitoring studies and/or improvement of symptoms with proton pump inhibitor therapy.

• Cyclic vomiting was defined on the basis of Rome III criteria as the presence of three or more discrete stereotypical episodes of vomiting in the prior year with absence of nausea and vomiting between episodes (4).

• Rumination syndrome also was based on Rome III criteria and was considered for those with persistent or recurrent regurgitation of ingested food not preceded by retching with subsequent remastication and swallowing (4).

• Postsurgical nausea and vomiting was characterized by symptoms of nausea and/or vomiting in the setting of prior gastric surgery and absence of delayed emptying.

• Medication-induced nausea and vomiting was distinguished by the resolution of nausea and/or vomiting with medication cessation or recurrence of symptoms upon rechallenge.

Numeric data were expressed as means or medians, and categorical data were expressed as proportions. Between-group comparisons were made using the chi-square or Fisher exact tests where appropriate.

RESULTS

Of 271 patients screened for inclusion, 23 were excluded: 8 had no record identifiable during the data collection period, and 15 were evaluated primarily for other complaints (6 for constipation, 4 for chronic diarrhea, 3 for abdominal pain, 1 for dysphagia, and 1 for allergies); none were pregnant or <18 years. A total of 248 patients were analyzed for the 3-year interval. Baseline characteristics are outlined in Table 1. The typical patient was female, middle aged, and had symptoms of nausea and vomiting for 2 years. Risk factors for delayed gastric emptying were ascertained. The majority had prior diagnostic evaluation that included endoscopy and abdominal imaging. Prior prokinetic medications were tried in 85%.

Table 1.

Baseline characteristics in 248 patients referred for chronic nausea and vomiting

Variable	Value
Gender
Women	201 (81%)
Men	47 (19%)
Age in years (median, range)	42 (18–78)
Symptom duration in years (median, range)	2 (0.2–50)
Symptoms
Nausea and vomiting	196 (79%)
Nausea alone	52 (21%)
Abdominal discomfort	145 (58%)
Prior hospitalization	98 (40%)
Risk factors for gastroparesis
Diabetes mellitus	63 (25%)
Prior gastric surgery	29 (12%)
Medications	91 (37%)
Prior evaluation
Endoscopy	196 (79%)
Abdominal imaging	135 (54%)
Neurologic imaging	34 (14%)

Of the total cohort, 156 (63%) were referred with a suspected diagnosis of gastroparesis. Of these, 102 (65%) had gastric emptying scan reports available for review, of which 95 (93%) were done with methods other than the 4-hour international standard protocol. Ninety (95%) of these had been interpreted as abnormal. A review of each study protocol revealed that the most common deviation from the international standard was an examination duration less than the recommended 4 hours (found in 43 [45%]). Other deviations included the use of alternative meals (in 7 [7%]), concomitant administration of medications known to alter gastric motility (in 2 [2%]), concurrent symptoms of nausea or abdominal pain during the test (in 2 [2%]), or a combination of these aberrations (in 41 [43%]). Repeat testing using the international standard protocol was done in 36 patients who had a prior abnormal nonstandardized GES; 27 (75%) of these repeat studies were normal.

History, examination, and targeted diagnostic evaluation revealed various diagnoses, as outlined in Table 2. Of the entire cohort, gastroparesis was confirmed infrequently despite the frequency with which patients had been referred with that diagnosis. This appeared to be largely the result of using nonstandardized GES.

Table 2.

Diagnoses in 248 patients referred for chronic nausea and vomiting

Diagnosis	n (%)
Chronic vestibular dysfunction	64 (26%)
Gastroparesis	28 (11%)
Cyclical vomiting syndrome	22 (9%)
Rumination syndrome	3 (1%)
Gastroesophageal reflux disease	5 (2%)
Postsurgical	6 (2%)
Medication-induced	3 (1%)
Other miscellaneous	41 (17%)
Unspecified	76 (31%)

Suspected CVD was the most common specific diagnosis, found in 64 (26%). Findings of altered vestibular function that led to this diagnosis included abnormal modified Fukuda stepping test (in 36 [56%]), presence of nystagmus (in 17 [27%]), and abnormal Romberg test (in 1 [2%]). Ten patients (16%) had a combination of these findings. Forty-three of these patients with CVD (67%) were referred for suspected gastroparesis. Of these, 34 (79%) had previous GES available for review. Thirty-three (97%) were done with nonstandard methods, of which 29 (88%) had been interpreted as abnormal. When a standardized test was repeated in nine of these patients, it was normal in eight.

Compared with patients with other conditions, CVD patients more often presented with nausea alone and tended to have fewer hospitalizations, suggesting less severe symptoms (Table 3). Abdominal discomfort and dizziness were not distinguishing symptoms, and overt vertigo was reported only in a minority of CVD patients (6 [9%]). To assess for possible contributing conditions, various comorbidities were evaluated; only eight CVD patients reported a recognized history of an inner-ear disorder.

Table 3.

Characteristics of patients diagnosed with chronic vestibular dysfunction versus other diagnoses among 248 patients referred for chronic nausea and vomiting

	Chronic vestibular dysfunction (n = 64)	Other diagnosis (n = 184)	P value
Gender
Men	10 (16%)	37 (20%)	0.42
Women	54 (84%)	147 (80%)
Symptoms
Nausea alone vs. vomiting	21 (33%)	31 (17%)	<.01
Abdominal discomfort	33 (52%)	112 (61%)	.20
Dizziness	28 (44%)	69 (38%)	.33
Vertigo	6 (9%)	2 (1%)	<0.01
Prior hospitalization	19 (30%)	79 (43%)	.06
Comorbidities
Inner ear disorder	8 (13%)	4 (2%)	<.01
Diabetes mellitus	13 (20%)	50 (27%)	.27
Prior gastric surgery	4 (6%)	25 (14%)	.17
Depression/anxiety	34 (53%)	75 (41%)	.09
Migraine	5 (8%)	27 (15%)	.15
Hypertension	15 (23%)	63 (34%)	.11
Atherosclerosis	5 (8%)	17 (9%)	.76
Thyroid disease	11 (17%)	24 (13%)	.41

Antivertiginous medication trials were administered to 57 of 64 (89%) CVD patients. These typically included scopolamine, meclizine, or benzodiazepines alone or in combination. Follow up was available for 39 of these patients for a median of 5 months. Symptomatic improvement was reported in 25 (64%), while 14 (36%) reported no change in symptoms.

DISCUSSION

This study has two main conclusions. First, despite the recommendations of experts to use a standardized protocol, most gastric emptying scans available for review were performed with nonstandardized methods. The most common deviation was duration less than the recommended 4-hour timeframe. Tougas et al (5) established a cut-off of >10% isotope retention at 4 hours after ingestion of a standard low-fat meal as indicative of delayed gastric emptying. Despite the validation of this cut-off as the international standard, many institutions still employ a variety of other protocols (6). When standardized exams were repeated at our institution, 75% were normal. Therefore, utilization of nonstandardized GES may lead to a misleading diagnosis of gastroparesis and should be abandoned.

Second, we found that CVD is a noteworthy consideration in patients with chronic nausea and vomiting, and treatments aimed at vestibular dysfunction may mitigate symptoms in some of those patients. CVD is classically characterized by symptoms of vertigo and postural instability but also may manifest with vegetative symptoms such as nausea and vomiting. Examination signs of CVD were found in 26% of our cohort. This is comparable to the 35% of adults over 40 years found to have CVD upon modified Romberg testing in the National Health and Nutrition Examination Survey (7). Although self-reported vertigo in population surveys has an estimated prevalence of 21% to 29% (8–10), the patients in our study presented with complaints of chronic nausea and vomiting and did not frequently report a history of vertigo or a known inner-ear disorder. The mechanisms whereby certain patients with CVD manifest symptoms are incompletely understood but may reflect impaired vestibular compensation to the initial underlying injury (11).

The suspected CVD diagnosis, in this study, was made most commonly on the basis of a positive modified Fukuda stepping examination. With visual, auditory, and proprioceptive inputs eliminated, patients must rely on their vestibular system to maintain orientation. Although the majority of normal subjects maintain orientation with 100 blindfolded steps in a quiet environment, Fukuda in his original experiments described a forward migration up to 1 meter and rotation up to 45 degrees from midline as normal (12). Patients with known vestibular lesions typically demonstrate >90-degree variations from midline (13). Therefore, to increase specificity for CVD, the 90-degree rotation was used as the cut-off in this study. Because advanced vestibular testing was not done as part of the evaluation of these patients, we cannot comment on the accuracy of the modified Fukuda test in this cohort of patients.

Therapy for CVD should ideally be targeted to the underlying lesion if identifiable, and affirmative physical examination signs should prompt referral for specialized confirmatory testing. Initial symptomatic improvement can occur with medications targeted at neurotransmitter targets within the vestibular system, such as antihistamines, anticholinergics, and benzodiazepines. The fact that nausea and vomiting improved with treatments aimed at vestibular dysfunction in many of our patients with suspected CVD supports the notion that CVD may play an important role in the pathogenesis of chronic nausea and vomiting in at least some patients with otherwise idiopathic nausea and vomiting. Future prospective studies should determine the frequency of vestibular dysfunction presenting with chronic nausea and vomiting, the efficacy of therapies such as medications and vestibular rehabilitation, and long-term prognosis.

In addition to CVD, other diagnoses were elicited in our patients suspected of having gastroparesis. These diagnoses included cyclic vomiting, rumination syndrome, gastroesophageal reflux, and postsurgical and medication-induced nausea and vomiting. This reinforces the heterogeneity of potential causes for nausea and vomiting that must be considered in the differential diagnosis. The frequencies of these conditions in our cohort at a tertiary referral center may not be representative of the frequencies in the general community; nevertheless, these are important diagnostic considerations for frontline gastroenterologists, as specific therapies may be more effective than nonspecific symptomatic therapy.

There are limitations to this retrospective study. First, complete prior diagnostic information was not available for review in all patients. Second, patients were not studied prospectively with a uniform diagnostic protocol. Ideally, a prospective study would include standardized GES, advanced vestibular testing, a uniform treatment protocol, and scheduled follow-up. Thus, the data collected are incomplete. Nevertheless, we feel that there are lessons to be learned from our experience.