False-positive head-impulse test in cerebellar ataxia

Abstract

The objective of this study was to compare the findings of the bedside head-impulse test (HIT), passive head rotation gain, and caloric irrigation in patients with cerebellar ataxia (CA). In 16 patients with CA and bilaterally pathological bedside HIT, vestibuloocular reflex (VOR) gains were measured during HIT and passive head rotation by scleral search coil technique. Eight of the patients had pathologically reduced caloric responsiveness, while the other eight had normal caloric responses. Those with normal calorics showed a slightly reduced HIT gain (mean ± SD: 0.73 ± 0.15). In those with pathological calorics, gains 80 and 100 ms after the HIT as well as the passive rotation VOR gains were significantly lower. The corrective saccade after head turn occurred earlier in patients with pathological calorics (111 ± 62 ms after onset of the HIT) than in those with normal calorics (191 ± 17 ms, p = 0.0064). We identified two groups of patients with CA: those with an isolated moderate HIT deficit only, probably due to floccular dysfunction, and those with combined HIT, passive rotation, and caloric deficit, probably due to a peripheral vestibular deficit. From a clinical point of view, these results show that the bedside HIT alone can be false-positive for establishing a diagnosis of a bilateral peripheral vestibular deficit in patients with CA.

Keywords: cerebellar atrophy; cerebellum; flocculus; head-impulse test; vestibuloocular reflex.

Figure 1

Mean peak slow phase velocity (SPV) for each of the 16 patients included in the study. According to the results of the caloric irrigation, the cerebellar patients are divided into two groups: patients 1–8 with mean peak SPV < 5 deg/s (CACR−) and patients 9–16 with mean peak SPV > 10 deg/s (CACR+).

Figure 2

Time course of the VOR gain during the head-impulse test. Boxplots show the VOR gain measured by the scleral coil technique at 40, 60, 80, and 100 ms for left- and rightward head-impulses for the group of patients with normal caloric response (CACR+) and reduced/absent caloric response (CACR−). Boxplot whiskers indicate the data range, the middle line the median value, and the edges of the boxes the upper and lower quartile. Although the initial VOR gain at 40 ms did not statistically differ between the two groups, gain values at 80 and 100 ms were statistically higher in the CACR+ group. Statistically significant values are marked with (*), for p values see text.

Figure 3

Head-impulse test recordings of a CACR− patient. Subplots (A,B,E,F) refer to rightward and subplots (C,D,G,H) to leftward HIT data. Subplots (A,C) show eye (red tracks) and head (blue tracks) velocity (deg/s) data for the first 250 ms after starting the head-impulse. Subplots (B,D) show the eye and head position (in deg) over time. Subplots (E,G) show the time course of the HIT gain (eye/head velocity). Subplots (F,H) show the position of the eye in space (in deg), here defined as position error (ideal value zero).

Figure 4

Head-impulse test recordings of a CACR+ patient. Subplots (A,B,E,F) refer to rightward and subplots (C,D,G,H) to leftward HIT data. Subplots (A,C) show eye (red tracks) and head (blue tracks) velocity (deg/s) data for the first 250 ms after starting the head-impulse. Subplots (B,D) show the eye and head position (in deg) over time. Subplots (E,G) show the time course of the HIT gain (eye/head velocity). Subplots (F,H) show the position of the eye in space (in deg), here defined as position error (ideal value zero).

Figure 5

Average latency of the first saccade made after initiation of the head-impulse for both groups. Data for right and left head-impulse are pooled. CACR+ patients make their first saccade almost constantly at about 200 ms (after the head movement is over), whereas CACR− patients make their first corrective saccade much earlier, often much before the head movement is over.

Figure 6

Average VOR suppression, VOR dark, and VOR light gain at 0.33 Hz for both groups. Although VOR suppression gains did not differ between the two groups, gains for middle frequency VOR in dark and in light were higher in the CACR+ than in the CACR− group, similar to the HIT gains VOR gains. Asterisks (*) indicate statistical significance between the two patient groups, error bars represent standard error of the mean.