Introduction

Cerebellar ataxia with neuropathy and vestibular areflexia syndrome (CANVAS) is an uncommon, late-onset form of ataxia characterized by the combination of bilateral vestibular, cerebellar and somatosensory impairment [1, 2]. Stage diagnostic criteria have recently been proposed, which for clinically definite CANVAS are as follows [3]: (i) abnormal visually enhanced vestibulo-ocular reflex (VVOR); (ii) cerebellar atrophy on MRI mainly involving vermian lobules VI, VIIa and VII; (iii) neurophysiological evidence of a sensory neuronopathy; and (iv) exclusion of genetic ataxias able to be gene tested. Several sets of sibling pairs have been reported, suggesting CANVAS is an autosomal recessive (AR) disorder [2, 4] (MIM#614575).

The somatic sensory deficit contributes to a significant level of disability in CANVAS [5, 6]. Neurophysiological studies have revealed an almost constant absence of sensory nerve action potentials (SNAP) in upper- and-lower limb nerves, and somatosensory evoked potentials (SEP) [5, 7]. The pathological background is a severe dorsal root ganglionopathy with secondary degeneration of central and peripheral sensory axons [4, 8]. Despite these electrophysiological and pathological features should be closely associated with generalized areflexia, intact muscle stretch reflexes have been reported in a significant proportion of CANVAS patients [1, 3]. The pathophysiology of such clinico-electrophysiological discordance remains unexplained.

Chronic cough has been described in association with several genetic disorders involving the autonomic nervous system, such as Holmes-Adie syndrome, Charcot-Marie-Tooth disease type 2J, dominant ataxia, and hereditary sensory and autonomic neuropathy (HSAN) [9,10,11,12]. Chronic cough and autonomic dysfunction may be an integral part of the clinical picture in CANVAS [2, 13].

The aim of this study was to report five CANVAS patients presenting with chronic, idiopathic cough, four of them coming from two non-consanguineous sibships. All patients showed severe somatosensory deficit and, leaving ankle reflexes aside, preserved muscle stretch reflexes and normal electromyographic deep tendon reflex responses (T-reflex). We postulate that, in spite of the absence of SNAP, in CANVAS patients with normoreflexia, functional muscle spindle afferents (Ia) are spared.

Patients and Methods

Patient data

The study is based upon five CANVAS patients, four of them belonging to two non-consanguineous pedigrees (Fig. 1), attended by us between 2006 and 2017. The severity of ataxia was graded according to IAPS (Table 1) [14]. The hospital ethics committee approved the study; all patients gave informed consent to participate in the study including photographs and video-recordings.

Fig. 1
figure 1

Pedigrees of Cases 1–4. a Patient number 1 (Table 1) is case III-5, and patient number 2 is case III-9. b Patient number 3 is case II-2, and patient number 4 is case II-1. Arrows indicate proband patients; horizontal bars indicate examined subjects; and slash line indicates died individual (ages at death are indicated under symbols). Squares are male and circles are female. Filled symbols are affected subjects and empty symbols unaffected subjects. Patients’ parents were not consanguineous

Full size image
Table 1 Summary of clinical data at last examination
Full size table

Electrophysiological study

Nerve conduction studies (median, ulnar, peroneal, tibial and sural nerves) were carried out as reported elsewhere [15, 16]. Biceps brachii T-reflex, patellar T-reflex, soleus T-reflex (Achilles T-reflex), and masseter reflex were recorded as reported by us [17, 18]. Needle electromyography (EMG) was performed analysing the duration and morphology of the motor unit potentials, the presence of spontaneous activity, and the electromyographic pattern at the maximal voluntary effort. Somatosensory evoked potentials (SEP) were performed according to standard methods [19].

Following standard methods, sympathetic skin response (SSR) was carried out while the subject was lying supine with the skin temperature controlled between 32 and 34 °C at the recording sites.

Laboratory investigations

All five patients were thoroughly investigated with screens for acquired causes of spinocerebellar degeneration including toxic, metabolic and systemic disorders (thyroid hormone levels, vitamin B12 deficiency, and serology for celiac disease and Sjögren’ syndrome and tumour markers). Molecular screening for Friedreich’s ataxia (FA) and spinocerebellar ataxia (SCA) (SCA1, SCA2, SCA3, SCA6, SCA7, SCA8, SCA12, SCA17 and DRPLA) was performed.

Imaging study

Non-contrasted magnetic resonance imaging (MRI) study of the brain was performed in all five patients.

Vestibular function testing

The diagnosis of bilateral vestibulopathy (BVP) was based on diagnostic criteria of the Classification Committee of the Bárány Society [20]. These include the following items: (i) the patient history of unsteadiness when walking or standing, which worsens in darkness, uneven ground or during head motion; (ii) no symptoms while sitting or lying down; (iii) bilaterally pathological horizontal angular vestibulo-ocular reflex (VOR) gain < 0.6 (at video head impulse test [vHIT]) and/or reduced caloric response (sum of bithermal maximal peak slow phase velocity on each side < 6°/s), and/or reduced VOR gain 0.1 upon sinusoidal stimulation on a rotatory chair (0.1 Hz, Vmax = 50°/s) and phase lead > 68 degrees (time constant < 5 s); and (iv) not better accounted for by another disease.

Results

Clinical findings

Clinical features of all five patients appear summarized in Table 1. Cases 1 and 2 were first cousins (Fig. 1a), and cases 3 and 4 (Fig. 1b) were sisters; case 5 was sporadic. There were 4 females and 1 male. Ages of patients ranged from 60 to 68 years (median 66) with gait instability initiated between 41 and 49 years (median 46).

Non-specific, diurnal, chronic and dry cough was the initial manifestation antedating gait instability from 15 to 29 years, excepting in case 4 who noticed the simultaneous occurrence of cough and gait disturbances (Table 1). Coughing usually came in bursts sometimes triggered by various factors: cold air, bending down or rising after sleep, eating bananas, dry foods (eg, popcorn, breadcrumb or dried fruit) (Supplementary material, Fig. 1) or spicy foods, inhalation of strong odours or dusts, strong speaking aloud, long telephonic conversations, and shouting. Neither stimulation the external auditory meatus with a cotton nor otoscopic procedures triggered cough reflex (Arnold’s ear-cough reflex) in any of the patients. The frequency of the cough bursts was variable over time, ranging from occasional throughout the year to almost daily presentation, duration varying between few seconds and several hours. None of the patients received cough-induced drugs (e.g., enalapril). Associated symptoms included scratchy and dry throat and voice changes, but none of the patients had dysphagia or syncope. Only patient 5 referred to symptoms of gastro-oesophageal reflux disease, which responded to lifestyle modification and anti-acid drugs with no changes to the frequency of cough bursts. Patients were evaluated by their general physicians and lung specialists without observing respiratory diseases or allergies; none of them did smoke. Several antitussives medications were administered in patient 1, just codeine phosphate being partially efficacious. Only in this patient, the coughing episodes tended to attenuate after the onset of ataxia. There were no other possible manifestations of dysautonomia, including persistent diarrhoea or constipation, sphincter disturbances, erectile dysfunction, dry eyes/mouth, sweating changes, or orthostatic intolerance.

All five patients showed postural imbalance and unsteadiness of gait, both of which worsened in darkness and on uneven ground, occasionally being accompanied by oscillopsia (Table 1). Patient number 1, a professional mountaineer and climber, noticed that his gait disturbances were initiated with difficulty to traverse narrow mountain passages by foot, obliging him to crawl out. None of them noticed hearing loss. All five patients showed marked vibratory hypopallesthesia up to iliac crests, and to a lesser degree in hands and forearms. There was positive Romberg sign and gait imbalance (Supplementary material, video) and stocking and glove tactile and pain hypoesthesia. Motor strength was preserved and neither amyotrophy nor fasciculations were noted. Upper-limb and knee jerks were either preserved or brisk, whereas ankle jerks were hypoactive or absent (Table 1); jaw jerk was systematically absent; and plantar responses, where obtainable, were flexor. At initial examination, three patients showed no appendicular cerebellar ataxia (cases 1, 3 and 5), though incoordination was noted upon serial examination (see video). Horizontal gaze-evoked nystagmus was noted in four patients and ataxic dysarthria in two of them (Table 1). Downbeating nystagmus was not observed. At the beside, the HIT test was systematically abnormal (see video). Two patients (1 and 5) referred to occasional lightning pains in thighs or lower legs.

Considering the prolonged clinical course (from 15 to 40 years, median 38), disease progression was slow with just one patient being wheelchair bound; IAPS stage ranged between 2 and 4, median 2 (Table 1). Video illustrates disease progression in patient 1, who in 2006, at age 53 and 20 years after onset, showed mild tabetic gait and postural imbalance, whereas 10 years later there was overt disease progression with worsening of tabetic signs, and appearance of cerebellar ataxia and dysarthria.

Serial evaluation of patient 3 was done over 6 years. At initial examination she showed pure postural rombergism, a working diagnosis of AR-HSAN being given, which was replaced by spinocerebellar syndrome with the appearance of appendicular ataxia and nystagmus, and subsequent evaluation of her affected sister (case 4) (Table 1).

Laboratory findings

Routine laboratory data, including those to exclude toxic, metabolic and systemic disorders, were normal or negative. No dynamic mutations associated with FA or SCA were found (see above for the list of SCA tests performed).

Electrophysiological findings

Table 2 summarizes the results of nerve conduction studies. Motor nerve conduction parameters, including F wave latencies, of median, ulnar, peroneal and tibial nerves were normal in every case. SNAPs of median and ulnar nerves were absent. In 3 out 4 examined patients, severely attenuated sural SNAP (between 0.2 and 0.5 µV) were recorded using needle electrodes (Supplementary material, Fig. 2), sensory conduction velocities being slightly slowed (between 34 and 39 m/s). SEPs from tibial nerve were absent in all four examined cases and absent or severely attenuated from median nerve (Supplementary material, Fig. 3).

Table 2 Results of nerve conduction studies
Full size table
Fig. 2
figure 2

Normal recordings of biceps brachii T-reflex (a case 3; b case 5) and Achilles T-reflex (c case 3)

Full size image
Fig. 3
figure 3

Cranial sagittal T1 weighted MRI images of case 1 (a, b) and case 3 (c, d). a Midsagittal view showing marked vermian atrophy mainly involving declive (VI), and folium vermis and tuber vermis (VII); note enlargement of the primary fissure (PF) and prepyramidal fissure (PPF). b Parasagittal view illustrating enlargement of the horizontal fissure (HF). c, d Similar but milder findings in case 3 exhibiting minimal cerebellar signs (Table 1). Cerebellar nomenclature is in accordance with Carpenter [21]. Note the absence of brainstem atrophy

Full size image

The biceps and patellar T-reflex was systematically normal, whereas Achilles T-reflex was absent in two cases (Table 1; Fig. 2). Electrophysiologically, masseter reflex was either absent or showed severely attenuated responses (Supplementary material, Fig. 4).

SSR, done in four cases (Nos. 1, 3–5), were normal.

MRI findings

Figure 3 illustrates cranial MRI findings observed in cases 1 and 3. In short, there was variable degree of vermian cerebellar atrophy mainly involving lobules VI, VII, VIIa, crus I and crus II [21] excepting in case 5 that showed normal findings; it is worth noting that this patient had minimal lower-limb appendicular ataxia and neither dysarthria nor nystagmus (Table 1).

Otoneurology investigation findings

All five patients exhibited abnormal bilateral video-HIT (VOR gain < 0.6) and caloric testing confirming (bilateral reduced caloric response).

Discussion

The current paper describes clinical data on 5 CANVAS patients presenting with spasmodic cough, whose clinical and electrophysiological examinations revealed preserved muscle stretch reflexes. Our series comprises two sisters and two first cousins coming from two unrelated pedigrees, a fact giving support to the notion that CANVAS is an AR syndrome [2, 4]. At last examination all 5 patients were in the seventh decade of life, their gait imbalance having been initiated in the fifth decade.

As reported in dominant ataxia with spasmodic cough [12], episodes of dry cough preceded gait instability between 12 and 29 years in four out of our cases. In this series a variable combination of factors precipitated cough bursts, particularly eating dry foods. Chronic cough has been reported in 2 out of 26 CANVAS patients, in one of them being the initial manifestation, 5 years before the onset of ataxia [13]. Szmulewicz and colleagues commented on that “we have found chronic cough and autonomic dysfunction to be variable features of CANVAS. Many patients report a relatively long-standing chronic, non-productive cough that may have preceded the onset of imbalance and somatosensory impairment” [2]. Although autonomic dysfunction has been described in the great majority of CANVAS patients [13], this was not the case in the current patients, in whom SSR were preserved suggesting functional integrity of non-myelinated postganglionic sympathetic fibres [22]. The pathophysiology of spasmodic cough in neurological disorders, including CANVAS, remains unsettled. It has been hypothesized that it may result from impairment of C fibre sensory innervation of either the upper airways or oesophagus, leading to denervation hypersensitivity of the secondary neurons in the nucleus solitarius [10, 11]. We did not observe that appearance of gait ataxia was correlated with strong attenuation of cough reflex, which would suggest that cough responsiveness is modulated by neural networks in the cerebellum [12]. Be that as it may, in cases of idiopathic, chronic cough the possibility of CANVAS should be added to the etiologic screening [23].

As characteristic of CANVAS, gait imbalance worsening in darkness was the presenting ataxic manifestation in this series [1, 2]. At initial examination, almost pure and severe rombergism in two cases led to establish a working diagnosis of HSAN. Serial examination was essential to monitor disease progression and assess the appearance of cerebellar signs (see Supplementary material, video). At last examinations all 5 patients exhibited severe lower-limb rombergism, less severe propioceptive deficit in upper limbs, preserved upper-limb and knee jerks, mild to moderate heel-knee ataxia, and positive head impulse test and vestibular areflexia on VVOR with video-oculography. Inconstant signs included upper-limb cerebellar ataxia, dysarthria, and horizontal gaze-evoked nystagmus. Progression of disease was slow, given that after several decades of onset, IAPS stage was severe in only one case. As a whole, these features are in line with those reported in CANVAS [1,2,3,4, 13].

Our nerve conduction studies revealed normal motor nerve conduction parameters, including F-wave latencies, and absence of SNAPs using standard surface recording electrodes. With needle recording in sural nerve severely attenuated SNAPs were obtained, resulting in slight slowing of the corresponding SCVs; these features could be accounted for by partial preservation of myelinated cutaneous afferents fibres [24]. SEPs from tibial nerve were absent, and absent or severely attenuated from median nerve. Our findings are in accordance with those reported by Szmulewicz and colleagues giving support to the notion that sensory neuronopathy, with central and peripheral axonopathy, is an integral part of CANVAS [1, 5].

In this series masseter reflex was electrophysiologically absent or with attenuated responses; these findings corroborate a previous study demonstrating absence or delayed responses in 8 out of 11 CANVAS patients [5]. Masseter reflex is unique among stretch reflexes, because its Ia fibres arising from muscle spindles have their bodies in the trigeminal mesencephalic nucleus in the brainstem, whereas the afferent cell bodies mediating other stretch reflexes are located in the dorsal root ganglia [25]. The masseter reflex may remain normal in patients with primary ganglionopathies even when stretch reflexes are abnormal in the limbs, as characteristically occur in FA; conversely, the masseter reflex is usually abnormal in axonopathies severe enough to cause areflexia of the limbs [26]. As proposed for FA and SCA2 [18, 27], in CANVAS there may be degeneration of mesencephalic nucleus accounting for such masseter areflexia [17, 27]; this calls for future autopsy studies with a systematic investigation of the central somatosensory system [28].

Notwithstanding the severe peripheral somatic sensory deficit and absence of SNAPs, tendon jerks were almost entirely preserved in our patients. T-reflex is an accurate tool for the assessment of nerve conduction in peripheral neuropathies and radicular syndromes [17]. With this in mind, we investigated T-reflex in 4 out of our 5 patients observing that biceps and patellar T-reflexes were systematically normal, as well as Achilles T-reflex in two of them. The presence of T-reflex implies normal function of both peripheral motor fibres and muscle spindle afferent fibres (Ia) [24]. Such features in the hallmark of a severe sensory neuropathy with central-peripheral axonopathy are very unusual. Let us focus on FA and HSAN type III (Riley-Day syndrome). As in CANVAS, classic FA phenotype is characterized by sensory neuronopathy with central and distal axonopathy implying areflexia and rombergism [29]. In late-onset FA with preserved muscle stretch reflexes, SNAPs and SCVs are either normal or slightly impaired, whereas SEPs are severely altered [27]. Therefore and unlike CANVAS, this FA subtype represents another cause of sensory neuronopathy with predominant central axonopathy [30], namely normorreflexia is correlated with preservation of the bulk of peripheral sensory axons. In HSAN III there may be complete absence of functional muscle spindle afferents, which could explain the loss of muscle stretch reflexes in a small-fibre neuropathy that, otherwise, would have them preserved [31]. Accordingly, we suggest that CANVAS is a unique syndrome with selective sparing of Ia fibres subserving muscle spindle afferents, in a setting of severe somatic sensory neuronopathy with widespread central-peripheral axonopathy [1, 5].

Our MRI studies corroborate a consistent pattern of cerebellar atrophy of anterior and dorsal vermis, mainly involving lobules VI (culmen), VII (folium vermis) and VIIa (tuber vermis) [1,2,3,4, 8, 32]. Laterally, there was marked enlargement of horizontal fissure indicating atrophy of the adjacent hemispheric lobules, crus I (superior semilunar lobule) and crus II (inferior semilunar lobule) [21]. This pattern of cerebellar atrophy was observed even in a patient with incipient cerebellar manifestations, namely MRI might be a good marker of the disease.

We conclude that spasmodic cough may be an integral part of the clinical picture in CANVAS, antedating the appearance of imbalance in several decades, and that sparing of Ia fibres is probably the pathophysiological basis of normoreflexia.