The efficacy of a modified Dix-Hallpike test with a pillow under shoulders

Abstract

BACKGROUND:

Dix-Hallpike (DH) test is sometimes done in a modified or reduced manner in a clinical setting. However, there has been no study that evaluated the reliability of this modified test.

OBJECTIVES:

The purposes of this study were to determine whether the efficacy of a modified DH test, named the ‘pillow under shoulders’ test, was equivalent to the standard DH test and to assess the difference in patient discomfort between the two methods.

METHODS:

Randomized controlled study at three academic referral hospitals, conducted in compliance with the CONSORT statement. Patients suspected for BPPV based on symptoms were randomly assigned to Group A or Group B. Patients in Group A received a standard DH test initially, followed by a modified DH test with a pillow under shoulders. Patients in Group B also received the two tests, but in the reverse order. The diagnostic results of both tests and patients’ subjective scoring for uneasiness (discomfort, pain, anxiety) were statistically analyzed.

RESULTS:

McNemar’s test and Kappa statistics showed a statistically equivalent diagnostic value between standard and modified DH tests (Cohen’s kappa = 0.823 and McNemar P = 0.18). The modified DH test had high sensitivity (95.5%) and fairly good specificity (87.9%). There was no statistical significance in the patients’ subjective scoring for uneasiness between the two methods, although most patients reported less inconvenience after the modified DH test compared to the standard DH test.

CONCLUSIONS:

The modified version of the standard DH test, the ‘pillow under shoulders’ method, may be a reliable, comfortable option to diagnose vertical canal BPPV. We suggest that this method can be used when head hanging is not feasible or sufficient for the standard DH test.

Keywords

Benign paroxysmal positional vertigo modified test Dix-Hallpike test vertigo

1 Introduction

Benign paroxysmal positional vertigo (BPPV), known to be a most common cause of dizziness, is characterized by short, repetitive vertigo induced by positional changes [4]. BPPV can occur in all three semicircular canals, and the posterior canal is most frequently affected. BPPV can occur as cupulolithiasis or canalolithiasis types, depending on the location of otolith in the semicircular canal. Canalolithiasis seems to occur more often, especially in posterior canal (PC) BPPV [9].

The gold standard diagnostic test for vertical canal BPPV is the Dix-Hallpike (DH) test [1]. A few alternative tests have been suggested for use of diagnosing posterior canal BPPV. In 2004, Cohen reported a ‘side-lying test (SLT)’ as a substitute for the DH test [2], and recently Michael et al. developed an ‘abbreviated posterior canalolithiasis chair-based assessment maneuver (APCCAM)’ as a chair-based version of the DH test [7].

In addition to the above alternatives, the DH test is sometimes done in a modified or reduced manner in a clinical setting. For example, the DH test may be done with limited neck extension or by making the patient lie on a pillow under shoulders. There has been no study that evaluated the reliability of these modified tests. Therefore, the authors aimed to determine whether the modified DH test, which is performed with a ‘pillow under shoulders’, is as reliable as the standard DH test. In this modified DH test, a pillow was placed under the patient’s shoulders when the patients reclined in the DH position, so that when the neck extends, the head rested stably on the table surface. The diagnostic efficacy of the modified DH test was statistically compared with the standard DH test through randomized controlled study design. The second objective was to investigate if there was a difference in the level of discomfort felt by patients after two kinds of DH test.

2 Participants and methods

2.1 Patients

The patients who visited three referral academic hospitals for repeated, brief periods of vertigo induced by head movement between February and August of 2013 were enrolled in this study. This study was a multicenter, prospective, randomized, single-blinded, and controlled clinical study. The study protocol was in compliance with the CONSORT 2010 statement for reporting randomized controlled trial, and the trial was conducted according to the Declaration of Helsinki and all its revisions. It was approved by the Institutional Review Board of the CMC Clinical Research Coordinating Center (IRB approval number: XC13OIMI0001 H). All patients provided written informed consent to participate in the trial.

Inclusion criteria were adults of 20 years of age or older with a symptom history suggestive of BPPV. We excluded patients who had been diagnosed and treated with BPPV previously in other hospital. Patients who had simultaneous neurologic or auditory symptoms, disease involving the central nervous system or cervico-lumbar vertebrae, severe obesity, pregnancy, or combined lateral canal BPPV were also excluded from this study.

2.2 Description of a modified DH test with ‘a pillow under shoulders’

During the standard DH test, the patient’s head is turned 45 degrees toward the testing side, and then the patient’s upper body and head are reclined backward with the head extended 30 degrees below the bed frame for approximately 1-2 minutes. The modified DH test, also referred to as a DH test with a pillow under shoulders, was done using the same direction and angle, but with a pillow placed under the patient’s shoulders, allowing the extended head to rest on the bed surface (Fig. 1). A dense foam pillow with 10-cm thickness was used for this maneuver. In a sitting position, the distance from hip to shoulder was measured in each patient and was used to determine the pillow placement. When DH test was performed with a pillow, the patient was laid on the bed with shoulders on the pillow, resulting in a natural neck extension over the pillow with the occiput resting on the surface of the bed. With this modified maneuver, the examiner only needs to guide the direction and angle of the patient’s movement and does not need to support the patient’s head at the last and most important step in the head-extended position. Because the patient’s head is now stably supported by the bed, patients are relived and are willing to cooperate with the procedure. The examiner does not need to support the patient’s head during head-extended position, thereby reducing the physical strain of the arm and back of the examiner, and allows them to concentrate on observing the patient’s eye movement.

Fig.1

Modified Dix-Hallpike (DH) test with a pillow under shoulders. (A) In a sitting position, the distance from hip to shoulder was measured. A dense foam pillow with a 10-cm thickness was placed on the bed at the same distance from the hip. (B) When the patient was laid back during the DH test, the patient’s shoulder and back rested on the pillow, which resulted in natural neck extension over the pillow, with the occiput resting on the bed surface. Because the patient’s head is now stably supported by the bed, patients felt comfortable and relived.

2.3 Experiment protocol

Using a random table, the patients were assigned to either Group A or B. Patients in Group A first underwent the standard DH test. After five minutes of rest to eliminate a fatigue effect, the modified DH test was performed. Patients in Group B performed the tests in the reverse order (Fig. 2). During the test, patients were wearing Frenzel glasses or infrared goggles, and eye movement was observed by the examiner. Latency, duration, and direction of nystagmus were recorded. All tests were conducted by experienced neurotologists (E-J-J, D-H-L, and J-H-S). After finishing each test, all patients were asked to evaluate the levels of discomfort, pain, and anxiety for each diagnostic maneuver on a visual analogue scale (VAS). The diagnostic criterion of vertical canal BPPV was presence of both characteristic nystagmus and subjective vertigo induced by the DH test. When the diagnostic criteria were fulfilled by either the standard DH or modified DH test, BPPV was diagnosed [9].

Fig.2

Study overview. A flow chart representing patient enrollment, random allocation, and two tests including a break for each group and questionnaire for the levels of discomfort, pain, and anxiety on visual analogue scale (0 to 10) after each test.

2.4 Statistical analysis

Statistical comparison of the two tests was performed using McNemar’s test, and the level of agreement between the two tests was assessed by Cohen’s kappa statistic, with values of 0.00 to 0.20 indicating poor agreement, 0.21 to 0.40 indicating fair agreement, 0.41 to 0.60 indicating moderate agreement, 0.61 to 0.80 indicating good agreement, and 0.81 to 1.00 indicating excellent agreement [5]. Sensitivity, specificity, and accuracy of the modified DH test and standard DH test were determined. The 95% confidence intervals were calculated using the exact method. Independent sample t-test was used to compare the demographic data of the two groups and the results of participants’ self-reporting score. Chi-square test and Mann-Whitney U test were used to compare the demographic data of the two groups. Statistical analysis was performed using SPSS version 18 (SPSS Inc., Chicago, IL, USA). All tests were two-sided and a P-value of <0.05 was considered significant.

3 Results

The total number of patients included in this study was 102, and 53 and 49 patients were assigned to Groups A and B, respectively, by a random table allocation. Among these, a total of 51 patients were diagnosed as vertical canal BPPV after standard and modified DH test (25 patients in Group A and 26 patients in Group B). Demographic data of all participants and patients with vertical canal BPPV in both groups are demonstrated in Table 1. All patients with vertical canal BPPV showed nystagmus corresponding with canalolithiasis, and no one showed nystagmus suggesting cupulolithiasis. There was no case with bilateral BPPV or multi-canal BPPV. Of the 51 patients with vertical canal BPPV, posterior canal type was suspected in 49, and anterior canal type was noted for 2.

Table 1
Demographic data of participants

Group A Group B P-value

All subjects

Number of participants 53 49

Age (year) 53.20±16.80 51.33±10.69 0.505

Gender (Male: Female) 13 : 49 15 : 34 0.492

Duration (day)^‡ 8.14±10.07 11.48±15.32 0.204

BPPV patients ^†

Number of patients PC-BPPV 23 26

AC-BPPV 2 0

Age (year) 54.40±14.60 53.00±8.94 0.684

Gender (Male: Female) 5 : 20 10 : 16 0.148

Duration (day) 5.56±5.57 12.60±17.57 0.062

Laterality Right 14 11 0.328

Left 11 15

		Group A	Group B	P-value
BPPV patients ^†
Number of patients	PC-BPPV	23	26
	AC-BPPV	2	0
Age (year)		54.40±14.60	53.00±8.94	0.684
Gender (Male: Female)		5 : 20	10 : 16	0.148
Duration (day)		5.56±5.57	12.60±17.57	0.062
Laterality	Right	14	11	0.328
	Left	11	15

^†BPPV patients: patients who were diagnosed in at least one of the standard and modified Dix-Hallpike tests. ^‡Duration: interval between the onset of symptom and the examination. (BPPV: Benign paroxysmal positional vertigo, PC: posterior canal, AC: anterior canal).

3.1 Diagnostic usefulness of the modified DH test

Among the 102 patients suspected for BPPV, the overall positive rate of standard DH test was 43.1% (44 from total 102 patients), and that of the modified DH test was 48.0% (49 from total 102 patients). McNemar’s test showed no statistically significant diagnostic difference between the standard DH test and modified DH test (p = 0.18), and the inter-test reliability for the two tests was found to be in excellent agreement, with a Cohen’s kappa of 0.82 (95% CI between 0.71 and 0.93). The sensitivity of the modified DH test was 95.5% (95% CI between 89.3 and 100%) and specificity was 87.9% (95% CI between 79.5 and 96.3%) (Table 2).

Table 2
Concordance and statistical analysis between the results of 102 patients examined by the standard Dix-Hallpike (DH) and modified DH tests

Modified DH test Standard DH test Total

Positive Negative

Positive 42 7 49

Negative 2 51 53

Total 44 58 102

Statistics Cohen’s kappa = 0.82 (95% CI between 0.71 and 0.93)

McNemar P = 0.18

Sensitivity of modified DH test = 95.5% (95% CI between 89.3 and 100%)

Specificity of modified DH test = 87.9% (95% CI between 79.5 and 96.3%)

Modified DH test	Standard DH test	Total
Positive	42	7	49
Negative	2	51	53
Total	44	58	102
Statistics	Cohen’s kappa = 0.82 (95% CI between 0.71 and 0.93)
McNemar P = 0.18
Sensitivity of modified DH test = 95.5% (95% CI between 89.3 and 100%)
Specificity of modified DH test = 87.9% (95% CI between 79.5 and 96.3%)

When comparing results according to test order regardless of test type, eight patients with a positive result at the first test were found to be negative at the second test, whereas one patient with a negative result at the first test was found to be positive at the second test. McNemar’s analysis was used to compare the diagnostic rate according to test order, which revealed that the first test had statistically higher diagnostic value than the second test (p = 0.039).

A separate analysis of each group yielded accordant result of no significant difference in diagnostic efficacy between the standard DH test and modified DH test (p = 1.00 and 0.07, respectively) with good or excellent reliability. (Table 3) When the modified DH test was used as the second test (Group A), the sensitivity of modified DH test was 96% and the specificity was 100%. When the modified DH test was performed as the first test (group B), the sensitivity of modified DH test was 94.7% and the specificity was 76.7%

Table 3

Separate analysis on the results of group A and group B

Group A (n = 53)				Group B (n = 49)
MDH test	SDH test (1st test)		Total	MDH test	SDH test (2nd test)		Total
(2nd test)	Positive	Negative		(1st test)	Positive	Negative
Positive	24	0	24	Positive	18	7	25
Negative	1	28	29	Negative	1	23	24
Total	25	28	53	Total	19	30	49
Cohen’s kappa = 0.96 (95% CI between 0.95 and 0.97)				Cohen’s kappa = 0.675 (95% CI between 0.646 and 0.704)
McNemar P = 1.000				McNemar P = 0.070
Sensitivity of MDH test as second test = 96.0%				Sensitivity of MDH test as first test = 94.7%
Specificity of MDH test as second test = 100%				Specificity of MDH test as first test = 76.7%

(MDH: modified Dix-Hallpike, SDH: standard Dix-Hallpike).

3.2 Patients’ subjective scoring for uneasiness

Although the survey analysis showed no statistically significant difference between the standard and modified DH tests in all three categories (discomfort, pain, anxiety), all participants and BPPV patients reported less discomfort, pain, and anxiety with the modified DH test compared to the standard DH test (Table 4). In terms of test order, the first applied test, regardless of standard or modified type, produced more discomfort, pain, and anxiety, although there was no statistical difference.

Table 4
The results of self-reporting score

Discomfort Pain Anxiety

Mean (SD) P-value Mean (SD) P-value Mean (SD) P-value

Total individuals Type of test

Standard DH test 2.72 (2.61) 0.17 1.26 (2.18) 0.84 2.17 (2.35) 0.49

Modified DH test 2.23 (2.48) 1.21 (2.00) 1.94 (2.29)

Order of test

First test 2.64 (2.60) 0.35 1.33 (2.22) 0.50 2.26 (2.51) 0.20

Second test 2.3 (2.50) 1.14 (1.96) 1.84 (2.11)

BPPV patients Type of test

Standard DH test 3.63 (2.76) 0.07 1.27 (2.05) 0.60 2.49 (2.58) 0.40

Modified DH test 2.69 (2.45) 1.08 (1.74) 2.08 (2.38)

Order of test

First test 2.64 (2.60) 0.18 1.33 (2.22) 0.41 2.26 (2.51) 0.22

Second test 2.30 (2.50) 1.14 (1.96) 1.84 (2.11)

		Discomfort	Pain	Anxiety
Total individuals	Type of test
	Standard DH test	2.72 (2.61)	0.17	1.26 (2.18)	0.84	2.17 (2.35)	0.49
	Modified DH test	2.23 (2.48)		1.21 (2.00)		1.94 (2.29)
	Order of test
	First test	2.64 (2.60)	0.35	1.33 (2.22)	0.50	2.26 (2.51)	0.20
	Second test	2.3 (2.50)		1.14 (1.96)		1.84 (2.11)
BPPV patients	Type of test
	Standard DH test	3.63 (2.76)	0.07	1.27 (2.05)	0.60	2.49 (2.58)	0.40
	Modified DH test	2.69 (2.45)		1.08 (1.74)		2.08 (2.38)
	Order of test
	First test	2.64 (2.60)	0.18	1.33 (2.22)	0.41	2.26 (2.51)	0.22
	Second test	2.30 (2.50)		1.14 (1.96)		1.84 (2.11)

^†BPPV: Benign paroxysmal positional vertigo, DH: Dix-Hallpike.

4 Discussion

BPPV is the most common peripheral vestibular disorder [1]. In most cases of BPPV, the cause is canalolithiasis of the posterior semicircular canal. When patients complain of brief positional vertigo, a positional test for PC-BPPV should be performed first. PC-BPPV is confirmed when characteristic nystagmus is induced by a DH test in patients with brief positional vertigo [8]. Through careful and exact performance of the test and deliberate analysis of the resulting nystagmus, the affected canal and side are well determined, and this leads to a successful canalith repositioning procedure.

Although the DH test is certainly the best to diagnose vertical canal BPPV, there are a few disadvantages and limitations to the DH test. First, patients need to passively recline their upper body and extend their head and neck into the intense vertigo-provoking position that is typically avoided. Further, patients must tolerate at least 30-seconds of head hanging supported only by the hands of an examiner, while withstanding vertigo. This inevitably causes severe fright and discomfort in the patient, regardless of the test result. Secondly, patients with any cervical spine or neck problem cannot participate in the test. Thirdly, because the patient’s head needs to hang 20° under the bed, the test cannot be done on hospital beds with a headboard; it is sometimes unavoidable for the patient to be moved to an examination table without a headboard. Finally, examiners may suffer arm pain and backache from holding onto the patient’s head while bending over for more than one minute.

Because of these limitations, alternatives to the DH test has been tried. Cohen suggested an SLT as a valid alternative to the DH test for use in patients with limited range of motion or who have difficulty relaxing [2]. In a side-lying position, the nose is turned 45 degrees away from the side being tested [2, 3]. Slow-phase eye velocity of nystagmus was not significantly different between the DH test and side-lying tests. Halker et al. extracted data from the Cohen study and analyzed sensitivity and specificity of the DH test and SLT. The sensitivity of SLT is higher than that of the DH test, while the specificity is similar between the two tests [3]. For the DH test, the estimated sensitivity was 79%, while specificity was 75%. For the side-lying test, the estimated sensitivity was 90%, and specificity was 75%. The disadvantage of SLT is that, for treatment with an Epley maneuver, the patient should change their place toward upper side of the bed. Otherwise, Semont maneuver can be used as canalith repositioning therapy in the same place of the bed as SLT, but the Semont maneuver needs rapid 180-degree rotation of head and body en bloc, which is difficult to achieve sufficient compliance in most patients [6].

In 2016, Michael et al. presented an abbreviated variation of the DH test and suggested that this test may serve as a screening procedure for quick identification of BPPV [7]. It is similar to the standard DH test, but is performed on a backed chair and does not require an examination bed/table. In the sitting position, the patient is leaned back against the back of the chair with the head turned 45 degrees toward the tested side, while the examiner guides the head of the patient as far back as possible. This chair-based simple version of the DH test had fairly good sensitivity (80%) and high specificity (95%). However, the authors considered positional nystagmus and/or positional vertigo/dizziness as a positive finding. A widely used diagnostic criterion of BPPV is the presence of both positional nystagmus and positional vertigo. However, if the universal criteria are applied, the observed sensitivity falls to 49.5%, while specificity reaches 100%. In the short version of the DH test, backward movement of the head is limited, so the head cannot be extended 20 degrees below the horizon, which is the most important step in the standard DH test for ampullofugal movement of the otolith by gravity. The relatively short movement distance and small angle in the short version of the DH test seem to be the reasons for its low sensitivity. Furthermore, 16 of 71 (18.4%) patients who do not show positional symptom or nystagmus were revealed to have posterior canal BPPV in the standard DH test.

In a clinical setting, BPPV patients experience anxiety during the DH test. During the test, they have to tolerate a passive, dizziness provoking position. In such cases, placing the patient’s head on the bed instead of hanging it in mid-air may comfort the patient during the examination, while also reducing the physical burden of the examiner. Also, many BPPV patients first visit the hospital through an emergency center, where the beds are usually equipped with a headboard; in such cases, the patients need to be moved to another examination table. Therefore, a reliable maneuver that can be implemented even on a bed with a headboard will be very useful. Although many clinicians have been performing a limited DH test that do not require neck extension and that place the patient’s head on a bed, there has been no reports on the resulting diagnostic rate or practicality.

For these reasons, the authors designed a modified DH test with a “shoulder on the pillow” technique, which places a pillow under the patient’s shoulders so that, when the neck extends, the head touches the bed surface. We compared the diagnostic efficacy of the modified DH test with that of the standard DH test, and found that there was no significant difference between the diagnostic rate of both tests and (43.1% versus 48.0%) and the inter-test reliability was in excellent agreement (Cohen’s kappa of 0.82). The sensitivity (95.5%) and the specificity (87.9%) of the modified DH test as a total (including first and second test) was quite high (Table 2). Therefore, in cases where a DH test needs to be done on a bed with a headboard, when a patient has a severe anxiety in the head hanging position or in a patient with limited range of neck motion due to neck or cervical spine problem, the modified DH test can be used as an efficient alternative. When the diagnostic rate of modified DH test was analyzed according to the sequence of the test, the sensitivity was consistently high regardless of the order of the test, while the specificity was low in the first test (Table 3). It is suggested that modified DH test can be used as a simple screening test because both primary and secondary examinations are highly sensitive. However, it is recommended to re-check with standard DH test if the initial modified DH test is negative because the specificity of the modified DH test is low during the first examination.

We measured patient’s subjective ratings such as discomfort, pain, and anxiety after each test and compared them. Although the difference was not statistically significant, the modified DH test yielded a lower score in all three items compared to the standard DH test.

A limitation of our study is that the second test was not performed with sufficient wait time until fatigue disappeared. While there was no difference in diagnostic rate between the two tests, second tests showed more false negative results; fatigability may be the reason of this.

Although there are some shortcomings, our randomized controlled study suggests that the modified DH test, named ‘pillow under shoulders’ test, is a valid test for diagnosing BPPV of the vertical canal. This modified test can be performed on the spot at an emergency room or hospital ward, where most beds are equipped with a headboard.

5 Conclusions

As the most common cause of dizziness, PC-BPPV can easily be confirmed with a DH test, and a simple substitute of the standard DH test in some difficult situations may be helpful in a clinical setting. The modified version of the standard DH test, named the ‘pillow under shoulders’ test, may be a reliable option that can be used when a standard DH test is not feasible. Since the specificity of the modified DH test is low during the first examination, it is recommended to re-test with standard DH test if the initial modified DH test results are negative.

Conflict of interest

The authors declare that there is no conflict of interest.

References

Bhattacharyya

, Gubbels

S.P.

, Schwartz

S.R.

, Edlow

J.A.

, El-Kashlan

, Fife

, et al., Clinical Practice Guideline: Benign Paroxysmal Positional Vertigo (Update), Otolaryngol Head Neck Surg 156(3_suppl) (2017), S1–S47.

Cohen

H.S.

, Side-lying as an alternative to the Dix-Hallpike test of the posterior canal, Otol Neurotol 25(2) (2004), 130–134.

Halker

R.B.

, Barrs

D.M.

, Wellik

K.E.

, Wingerchuk

D.M.

and Demaerschalk

B.M.

, Establishing a diagnosis of benign paroxysmal positional vertigo through the dix-hallpike and side-lying maneuvers: A critically appraised topic, Neurologist 14(3) (2008), 201–204.

Korres

S.G.

and Balatsouras

D.G.

, Diagnostic, pathophysiologic, and therapeutic aspects of benign paroxysmal positional vertigo, Otolaryngol Head Neck Surg 131(4) (2004), 438–444.

Landis

J.R.

and Koch

G.G.

, The measurement of observer agreement for categorical data, Biometrics 33(1) (1977), 159–174.

Mandalá

, Santoro

G.P.

, Asprella Libonati

, Casani

A.P.

, Faralli

, Giannoni

, et al. Double-blind randomized trial on short-term efficacy of the Semont maneuver for the treatment of posterior canal benign paroxysmal positional vertigo, J Neurol 259(5) (2012), 882–885.

Michael

P.C.

, Oliva

, Nunez

, Barraza

, Faundez

J.P.

and Breinbauer

H.A.

, An Abbreviated Diagnostic Maneuver for Posterior Benign Positional Paroxysmal Vertigo, Front Neurol 7 (2016), 115.

Shim

D.B.

, Song

C.E.

, Jung

E.J.

, Ko

K.M.

, Park

J.W.

and Song

M.H.

, Benign paroxysmal positional vertigo with simultaneous involvement of multiple semicircular canals, Korean J Audiol 18(3) (2014), 126–130.

von Brevern

, Bertholon

, Brandt

, Fife

, Imai

, Nuti

, et al., Benign paroxysmal positional vertigo: Diagnostic criteria, J Vestib Res 25(3-4) (2015), 105–117.