Slow Orthostatic Tremor: Review of the Current Evidence

Historical perspective

The first clinical and electrophysiological description of OT was by Pazagglia et al. in 1970. They reported three patients with leg tremor and unsteadiness emerging upon standing and relieved by leaning, ranging from 10 to 16 Hz on surface electromyography (EMG).³ Fourteen years later, Heilman coined the term “orthostatic tremor” to clinically describe three similar cases.^4,5 The term “slow OT” was first used by Thompson et al. in 1986, with electrophysiologic capture of both slow (8 Hz) and high (16 Hz) frequency tremors in a patient.⁶ In the same year, slow OT was described in a family with essential tremor (ET).⁷ Other observations emerged of cases with either classical or slow OT with coexistent ET, of which some responded to ET medications.⁸ It was proposed that OT was an ET variant, regardless of frequency.^8–10 although others argued against this.^11,12 Meanwhile, other reports emerged of lateralized slow OT occurring prior to or coincident with a diagnosis of parkinsonism, of similar slow frequency to coexistent upper limb rest tremor, and with exquisite responsiveness to levodopa, suggesting it was a parkinsonian variant.^13–17 In 2007, Thomas et al. proposed “pseudo-orthostatic tremor” to describe slow OT affecting both legs, with dramatic levodopa response, to differentiate it from parkinsonism-related OT and classical OT.¹⁸

In the last 5 years, further divisions of slow OT were proposed. In 2014, Erro et al. expanded the clinical and electrophysiological description of OT, and recommended the term “pseudo-orthostatic tremor” for all OT <13 Hz to reflect differences from primary OT.¹⁹ In 2015, Rigby et al. described the first series of slow OT patients divided into low (<10 Hz) and intermediate (10–13 Hz) frequencies, and contrasted with primary OT (>13 Hz). The clinical features of patients with low and intermediate frequencies differed from primary OT, whereas the electrophysiological features of intermediate frequencies were more closely aligned to high frequencies.²⁰ In 2016, Ure subdivided slow OT into three subtypes: “primary” (i.e., idiopathic), “secondary” (i.e., symptomatic), and “pseudo-orthostatic” tremor (asymmetry, parkinsonism, and levodopa-responsive).²¹ This brings us to the current timeframe, where the Consensus Statement took a new direction, again reducing all orthostatic tremors <13 Hz to the term “pseudo-orthostatic tremor.”¹

There is obvious widespread disagreement (nomenclature, classification, proposed pathophysiology) with unresolved questions and contradictory ideas. Should all tremors under 13 Hz be considered a single entity? If so, what is the optimal terminology? Why not simply “slow OT” to reflect the lower frequency if they are to be lumped together? The term pseudo-orthostatic tremor might be misinterpreted as functional tremor, or that patients were not truly orthostatic or weightbearing at the time of tremor; should this term be best reserved for re-emergent parkinsonian leg tremor while standing? Electrophysiology is currently the gold standard diagnosis, but is not accessible to many neurologists. Therefore, are there clinical clues that could favor slow OT over other mimickers of shaky legs? Should it be subdivided into parkinsonian variant, ET variant, slow variant of classical OT, ataxia variant, or other variant based on clinical and electrophysiological features?

We thus aim to methodically review the literature on slow OT to assess whether some of these questions can be answered, and to propose strategies to address unresolved issues. For the purposes of this article, we will use the term “slow OT” to reflect any orthostatic tremor less than 13 Hz.

Clinical

Of the 70 slow OT cases, two-thirds were female (44F; 26M; 63%), of mean age 60 years (range 26–86), and mean disease duration of 6 years (range 0–32). These demographics are similar to classical OT.²⁶ Almost all patients reported tremor or unsteadiness provoked by standing, that disappeared with walking, and was absent with sitting and lying. Slow OT persisted walking backwards in one patient,²⁷ although this aspect was not studied by others. It could worsen when walking slowed down.⁶ Rigby et al. noted that gait unsteadiness, falls, and abnormal gait were much more common in low (<10 Hz) compared with intermediate (10–13 Hz) OT, but these features were uncommon or absent in high-frequency OT (>13 Hz).²⁰ One-third of slow OT cases (23/70; 33%) occurred in isolation, challenging previous views that suggested it was rarely isolated.^28,29 Two-thirds of cases (n = 47 patients; 67%) had a coexistent disorder to which slow OT was attributed (see Table 3). These were almost all neurologic (n = 43); parkinsonism (n = 21; 30%), cerebellar ataxia (n = 8; 11%), dystonia (n = 7; 10%), and spasticity/myelopathy (n = 6; 9%) were most commonly identified, and occasionally more than one were concurrent. Autoimmune (n = 4; 6%), drug-induced, polyneuropathy (all n = 2; 3%) and aqueduct stenosis with hydrocephalus (n = 1; 1%) were uncommon. Within the autoimmune category, two patients had Graves’ disease. ET was formally diagnosed in only two patients (3%). However, a coexistent tremor was observed or recorded in 42 patients (60%): postural arm tremor in 24 (34%); rest tremor in 10 (14%); mixed postural and rest tremor in 5 (7%); postural leg; lips; or voice (all 1 each). Side-locked resting tremor and lateralized orthostatic tremor were noted in three patients with parkinsonism, highly suggestive that slow OT represented re-emergent leg tremor upon standing in these cases.³⁰ Other PD patients had resting tremor in the upper limb only, with a crossed orthostatic tremor in the opposite limb. Some PD patients had bilateral orthostatic leg tremor, without a resting leg component identified.

Imaging

There were focal structural and/or functional imaging abnormalities in about one-fifth of cases (15; 21%). This may be an underestimate, given that several cases reported were prior to widespread availability of magnetic resonance imaging (MRI). MRI brain or spinal cord images were abnormal in all patients with a coexistent structural neurological disorder, most commonly reflecting lesions in the cerebellum, pons, medulla, and upper spinal cord. In a patient with a lateralized pontine lesion, slow OT was more pronounced ipsilateral to the lesion. Single-photon emission computerized tomography (SPECT) imaging was abnormal in several cases with coexistent parkinsonism. In the remainder, brain imaging was normal or showed only diffuse age-related changes.

Neurophysiology

The median tremor frequency of slow OT was 6–7 Hz and ranged from 3 to 12 Hz. The tremor bursts ranged from 50 to 150 ms duration, and were often variable.^17,27,31 The bursts could be either synchronous^7,14,27,28 or alternating^6,32,33 in analogous muscles (e.g., bilateral tibialis anteriors); and either synchronous^34,35 or alternating in agonist/antagonist muscle pairs^8,18 (e.g., ipsilateral tibialis anterior and gastrocnemius). Low coherence (0.2–0.8) was reported,²⁰ and several cases with high coherence (≥0.8) were also identified.^18,27,36 These neurophysiological findings can be contrasted with classical OT, which is foremost immediately distinguished by its higher tremor frequency. Furthermore, classical OT typically has synchronous EMG tremor bursts in analogous leg muscles (e.g., bilateral tibialis anteriors) of uniform short duration (25–40 ms).^13,16 This synchronous bursting pattern has been reported in leg, paraspinal, and upper limb muscles where patients bear weight, or where tremor is activated by isometric contraction. Classical OT was previously regarded as the only organic tremor with high coherence (0.8–1.0), suggesting a single or closely linked bilateral oscillators, which distinguishes it from other centrally generated tremor disorders.³⁷ For example, essential and parkinsonian tremors have poor side-to-side coherence, suggesting discrete bilateral generators for these centrally generated movements.³⁸

Slow OT can be distinguished electrophysiologically from its mimic of orthostatic myoclonus. While orthostatic myoclonus can fall within the frequency range of slow OT, it can also occasionally be >13 Hz. However, it is distinguished from slow orthostatic tremor by an irregular nonrhythmic bursting pattern, and the burst duration is typically much shorter than slow OT, at under 100 ms and often <50 ms.³⁹

Rigby et al. compared electrophysiology in slow OT patients, divided into <10 Hz (n = 8) and 10–13 Hz (n = 6), versus classical OT >13 Hz (n = 14). Lower frequency discharges tended to have a broader spectral peak, greater variability in discharge duration, and lower intermuscular coherence.²⁰ In another study, patients were subdivided into “slow” (4–6 Hz) and “intermediate” (7–9 Hz) OT, and intermediate OT has shared EMG characteristics with “slow” and “fast” OT.^17,27

Slow OT has been observed to be evoked by different conditions of strong tonic-muscle contraction, such as independent standing, isometric muscle contraction, or pressing a limb against resistance.^7,28 This has also been observed with classical OT.^40,41 However, as it is sometimes not purely orthostatic or weight bearing, some have suggested reserving these terms for those characteristics.^7,39

Focusing on the electrophysiology findings of PD patients with resting leg tremor and orthostatic tremor, in three cases there was asymmetric or sidelocked tremor bursts on standing, suggestive of re-emergent leg tremor. However, in several other PD cases with rest leg tremor, the electrophysiology reports were lacking in details, and did not specify whether there was laterality of standing tremor.

Differential diagnosis

This is listed in Table 4. Slow OT can clinically mimic other shaky leg disorders, for example, classical OT, orthostatic myoclonus, or functional shaky legs, because the frequency and regularity of the shaking are difficult to accurately estimate by clinical examination. The “helicopter sign” of auscultating the legs with a stethoscope for a rhythm is not specific to orthostatic tremor.⁴² Electrophysiology remains the gold standard detection, confirming the presence of tremor bursts and the slow frequency.³⁹ Given limited access to electrophysiology studies in some clinical settings, emerging use of new technologies, such as app-based tremor devices which have been found to be highly sensitive in OT, can increase bedside detection.⁴³

Treatment

There are three main medication classes to which patients with slow OT appear to respond in a dichomotous fashion: parkinsonian medications (levodopa, anticholinergics), ET medications (beta-blockers, primidone), or classical OT medication (clonazepam). In addition, the two cases with Graves’ disease had a dramatic improvement after methimazole therapy to treat the underlying condition. These observations support earlier theories that slow OT has different variants with different pathophysiological underpinnings. Clonazepam and levodopa were the most effective medications (both n = 15; 21%), followed by gabapentin (n = 6; 9%) and primidone (n = 5; 7%). However, important observations can be made. Clonazepam and gabapentin were effective across all tremor frequencies. Levodopa was effective in 15/19 patients with 4–7 Hz OT, and 11 of these 15 responders had coexistent parkinsonism. Primidone was effective for 5–10 Hz tremor. Moreover, these findings suggest patterned variants of slow OT, characterized by frequency, coexistent disorder, and medication response. These medications may be helpful in classical OT, excepting levodopa which is often ineffective.²⁶ A variety of other medications were tried with mostly unfavorable responses.

Pathophysiology

There are several lines of support for slow OT having a central generator like classical OT, rather than of peripheral origin.⁴⁴ First, it can share the same frequency as concurrent tremor in other body parts, suggesting a single generator.³² The tremor cannot be reset by peripheral stimuli.^6,28,45 Stretch reflexes in the calf and pretibial muscles are typically normal.⁶ Peripheral neuropathic processes are rarely reported. Where studied, sensory nerve conduction studies and sensory and motor evoked potentials have been normal.^6,28

A combination of fast and slow orthostatic tremors were reported in five cases. Interestingly, three of these had a combination of 8 and 16 Hz OT, and subjective unsteadiness and visibly increased tremor amplitude was reported only during the presence of the 8 Hz tremor.^6,15,40 In one patient, selective slowing to 8 Hz was observed in one leg muscle, while remaining at 16 Hz in the other leg muscles.⁶ With treatment, the 8 Hz tremor disappeared and symptoms improved in two patients, while the 16 Hz tremor persisted.^6,15 These observations are intriguing and provide further data for speculation on the pathophysiology. In some patients, dual slow and fast frequencies have been attributed to either fast harmonics in slow OT⁴ or to slow harmonics in fast OT, depending on the predominant tremor frequency on accelerometry. However, the tremors in those three cases did not behave like a harmonic, as only one tremor resolved with treatment. Furthermore, the selective slowing of frequency in one leg muscle might suggest two distinct generators for fast and slow tremors, which may become coupled or uncoupled by an enigmatic mechanism. An alternative hypothesis is that they may have the same central generator, with frequency output modified by a secondary process, which is amenable to medication effect. Patients with dominant low frequency orthostatic tremor accompanied by subpeak harmonic or nonharmonic higher frequencies have been observed, and it has been theorized that other common frequencies might drive the lower limb tremor.^27,31,36

A shared underpinning for slow OT is suggested in patients with coexistent neurologic disorders manifesting with tremor of similar frequency as slow OT. A similar rationale can be made for structural lesions^13,25,26 or autoimmune disorders^13,29,33 that appear to be timelocked to the onset or offset of slow OT. This is aptly illustrated by the two patients with Graves’ disease, both with 8 Hz tremor, with complete remission after treatment with methimazole, suggesting the slow OT mechanism was due to enhanced physiological tremor. Interestingly, others have suggested that slow OT may be an exaggeration of physiological tremor, due to synchronization of central structures involved in organizing postural signal, for example, spinal cord neurons, inferior olive, or thalamus, which are all capable of spontaneous oscillatory activity.^{20,28,33,36,46,47} Central pathways involved in relaying postural signals could also be activated by standing, and trigger a central oscillator for OT.^14,28,40 The cerebellum, pons, medulla, and upper spinal cord have been hypothesized as potential sites based on symptomatic cases of slow OT.^34,46,48,49 However, as OT is a rare complication of structural brain and cord abnormalities, possibly a second hit in an intersecting pathway may be required for this phenomenon to occur.