Pharmacological treatment in Parkinson's disease: effects on gait

Straight walking

The hallmark motor symptom of PD, bradykinesia, is reflected in gait. The slow gait associated with PD is predominantly due to short steps, rather than slow cadence [19, 20]. Short steps are caused by slow, weak force development to push-off the foot [21, 22]. Short steps may also reflect imbalance as it is associated with higher double support time, i.e. both feet are on the ground for a longer proportion of the step cycle. Short steps may thus reflect an inability to control body CoM on one leg during a long step. Patients with PD are capable of modulating cadence, thereby at least partly compensating for small step length [20, 23]. In the upper body, bradykinesia is expressed by reduced arm swing amplitude and velocity, particularly in the most affected arm, and decreased range of motion of the trunk [4, 24].

In addition to bradykinesia of gait, PD gait is characterized by postural instability. Postural instability during gait can be observed in larger step-to-step variability compared to healthy subjects [1, 25]. For example, the duration, length, and width of steps vary more over multiple steps in people with PD than their healthy peers. This variability is believed to be related to increased fall risk in PD [26, 27], although this could not be confirmed in a large prospective study in PD [28].

Gait initiation

Subjects with PD show hypometric APAs when initiating a step. They produce less force, resulting in a smaller weight shift forward and toward the stance leg than would be required for taking a step [21, 29]. This reduction in push-off force causes a delay in step execution [21, 30]. Moreover, the length of the first step is shorter than in healthy elderly subjects [29, 31-33]. In contrast to healthy people, people with PD do not scale the APA size in proportion to their stance width [33]. Unlike the spatial control of APAs, the relative timing (coordination of the movement pattern) is unaffected by PD [30].

Turning

The sequential eyes-head-trunk-feet movement observed during turning in healthy adults is lost in PD patients; they turn “en bloc”, with a more simultaneous onset of eyes, head, trunk and leg movement [34-37]. Speed of turning in people with PD is related to axial tone in the neck, but not trunk or hips [38]. In addition, turning in people with PD is characterized by slow, jerky turning and additional steps are needed to complete the turn [39-41]. When people with PD are asked to turn quickly, they become unstable as they spend more time than control subjects with their body CoM outside their base of foot support [15].

Gait adaptability

Problems stepping over an obstacle while walking are apparent in people with PD who have moderate (H&Y 2-3) disease severity [42-44]. People with PD have a higher number of obstacle hits compared to healthy subjects [44, 45]. During obstacle crossing, people with PD are less stable than healthy subjects, reflected in more mediolateral movement of the CoM [46]. In addition, breakdown in motor planning can result in riskier situations. For example, the time needed to step over the obstacle is longer, resulting in more time spent standing on one leg [44, 47]. Also, the typical foot placement relative to the obstacle is smaller in PD patients than in healthy subjects, increasing the risk of collision [42, 44, 45].

Freezing of gait

A majority of people with PD eventually develop freezing of gait (FoG) – the subjective feeling that the feet are ‘glued to the ground’ – during the course of the disease [48]. Even without a freezing event, patients prone to FoG show differences in gait parameters compared to PD patients who do not freeze. People with PD and FoG walk more slowly with shorter stride lengths than non-freezers [49, 50], which could be due to more advanced disease stage in FoG [7]. Nevertheless, FoG has been associated with greater spatial stride-to-stride variability [50, 51], more time spent in double support phase [49, 52], and increased asymmetry between the right and left legs [52]. During turning, subjects with PD who have FoG require more time to turn and have a smaller step width than those without FoG [53].

All abovementioned aspects of gait impairments typically call for more attentional control of walking [54]. A loss of automaticity of gait can be reflected in more deterioration of gait while performing a secondary task (dual task cost). In people with PD, increased dual task costs in comparison to age-matched control subjects are most prominent in gait stability measures, and in people with more severe gait impairments, such as FoG [1, 55, 56].

Straight walking (Table 1)

Similar to improvements in other measures of bradykinesia, gait parameters that are related to movement amplitude and speed benefit from dopaminergic treatment. For example, increases in the amplitude of the movements of the legs, trunk and arms result in larger step length, trunk motions, arm swing [24], and faster stride velocity [59-62]. In contrast to spatial gait metrics, temporal metrics are not improved as much by levodopa. Cadence improves to a lesser extent [59], or not at all [61]. There is no evidence for decrease in double support time, which might be expected if levodopa reduced imbalance. However, and contrary to the evidence above, one study suggests that variability of step time, stride length, and gait speed can improve with dopaminergic treatment [61]. Some evidence also suggests that asymmetry of arm swing movements benefit from dopaminergic therapy [24].

Gait initiation (Table 2)

Levodopa can improve the hypometria of the preparatory stepping phases. Force generation for push off increases, resulting in higher amplitudes of the postural preparation phase (APA [21, 33, 59]) and larger first steps [33]. In contrast, neither the duration of the APA, nor the size of the first step was improved by levodopa in the study of Curtze et al. [59].

Turning (Table 2)

Whereas turning in place does not seem to benefit from dopaminergic treatment [38, 63], making a turn to change walking direction does. Turning speed and number of steps to complete the turn improved with levodopa [38, 59], but to a lesser extent than parameters of straight walking [59]. Interestingly, some evidence suggested that only patients with dyskinesia improved turning speed on dopaminergic medication [59].

Gait adaptability (Table 2)

Two studies evaluated the effect of dopaminergic treatment on gait adaptability. The ability to step over an obstacle during walking improved when using dopaminergic medication [45]. Success rates at avoiding obstacles changed in the ON state, with shorter and faster steps to cross the obstacle, suggestive of less cautious behavior [45, 64].

Freezing of gait

In most patients with PD and FoG, FoG episodes are more frequent, and last longer, when they are off dopaminergic medication compared to the on state [51, 65, 66]. In sharp contrast, a small group of people with PD have FoG that is induced by levodopa [67]. The ELLDOPA trial showed that FOG severity could be reduced or delayed when on dopaminergic treatment [68]. However, there is very limited evidence that administration of levodopa and other dopaminergic agents can ameliorate stride length and stride time variability in FoG to levels approaching non-freezers ([52, 65, 69]).

Dopamine Agonists

Dopamine D2 agonists are commonly used in PD, either as monotherapy in early PD, or to reduce levodopa dose needed in more advanced disease stages [70]. The specific effect of dopamine agonists as monotherapy has been evaluated in an open-label uncontrolled study with previously-untreated subjects with PD. In this study, six month use of the transdermal patch rotigotine improved all aspects of gait compared to baseline, including straight walking, gait initiation and turning [71].

The additional value of DA agonists as an adjunct to levodopa has been evaluated in three studies. Two studies found beneficial, immediate effects on gait speed of the DA agonist apomorphine (sublingual) and pramipexole compared to levodopa alone [72, 73]. In a large-scale, placebo-controlled study, the long-term effect of pergolide, in combination with levodopa, yielded better effects on the gait item of the UPDRS-III than placebo or levodopa alone [74]. Together, these studies suggest that dopamine agonists augment the effects of levodopa in advanced PD.

Catechol-O-Methyl Transferase (COMT) inhibitors

The effects of COMT inhibitors (entacapone, tolcapone) are similar to DA agonists, extending the effects of levodopa therapy. COMT inhibitors inhibit the methylation of levodopa, thereby increasing the concentration of levodopa that enters the brain. As such, COMT inhibitors, in conjunction with levodopa, are recommended to decrease the amount of OFF time. Only two small studies looked at the effects of COMT inhibitors on gait parameters, providing support for tolcapone as an effective add-on to levodopa to prolong beneficial effects on gait speed [75, 76].

Monoamine oxidase type B inhibitors (MAOB-I)

MAOB inhibitors (selegiline and rasagiline) have been evaluated in people recently diagnosed with PD, in two large-scale studies, with the aim to modify disease progression: DATATOP [77] and ADAGIO [78]. Although gait was not an outcome in the original study, results from a follow-up study suggested that PD patients on MAOB-I were less likely to develop FoG [79, 80] and had less increase of PIGD scores over a period of 36 weeks [81]. These findings were in line with earlier work showing marginally faster gait speed in PD subjects on MAOB-I compared to placebo [82]. However, a more recent, retrospective study could not replicate a beneficial long-term effect of MAOB-I on FoG or fall risk [83]. In addition, clinical lore suggests that addition of a MAOB-I to a patient developing FoG does not reduce the symptom. Future work to evaluate the effects of MAOB-I could benefit from the use of more objective and sensitive measures of gait.

Amantadine

Amantadine, an older antiviral agent, is a glutamate antagonist recommended for the treatment of dyskinesia in PD [70]. Recently, amantadine has been tested for its potential ameliorative effect on FoG in PD. Although there was initially some promise for short-term benefits for FoG from an uncontrolled trial [84], two double-blind studies failed to see improvement of FoG in people with PD after several days of amantadine use when compared with placebo [85, 86]. However, a double-blind, placebo-controlled study reported some improvements on secondary outcome measures of FoG, such as whether or not subjects reported FoG on a questionnaire [85]. In addition, amantadine has been tested for PD patients with axial impairments after STN-DBS surgery [87]. This uncontrolled study reported improvement on the gait item of the UPDRS-III. To conclude, more long-term and well-controlled studies are needed to confirm a potential effect of amantadine on objective gait measures.

Glutamatergic agents

Apart from amantadine, memantine is a glutamate agent acting as N-methyl-D-aspartic-acid-related (NMDA) receptor antagonists that has shown promising gait benefits in demented subjects [88]. There is one (phase I) study in PD evaluating the use memantine as adjunct-therapy to levodopa in patients with severe gait and balance impairments [89]. This pilot study did not report improvements in gait speed or stride length in the group on memantine compared to placebo. However, axial signs as assessed in the UPDRS-III did show some improvement.

Cholinergic medication

Recent research has highlighted the role of cholinergic activity in gait and balance impairments in Parkinson's disease [90-93]. Loss of cholinergic activity of the PPN-thalamus tract in PD has been related to fall risk [91, 94], and reduced cholinergic activity in the cortex associates with slower gait speed in people with PD [90]. This could suggest that loss of cholinergic activity in the cortex reduces gait speed through attentional loss, in line with consistently reported associations between attention and gait parameters [1, 28, 95, 96].

Cholinergic treatment with cholinesterase inhibitors (ChE-I; for example rivastigmine, donepezil) is commonly used for cognitive enhancement in patients with Alzheimer's disease (AD), and has also shown efficacy in patients with Parkinson's disease dementia [97]. Preliminary findings suggested that ChE-I can reduce falls in people with PD at high risk for falling [98] and improve gait in AD [99-101], potentially through improving attention. A recent, large double-blind, placebo-controlled trial (RESPOND trial [102]) evaluating gait variability and, secondarily, fall risk provided additional support for the beneficial effects of rivastigmine on gait for PD patients at risk for falls. Another phase II study is currently underway to establish the effect of donepezil on gait, postural stability and fall risk in PD, and evaluate its effect on attentional demands of gait [103].

Norephinephrine agents

In the early 1980s, a number of studies in Japan were conducted to evaluate the effect of L-threo-3,4-dihydroxyphenylserine (L-DOPS), a precursor of norepinephrine, on FoG in PD. Two open-label, uncontrolled studies reported conflicting results on FoG through L-DOPS administration [104, 105]. More recently, the a study on the effect of L-DOPS in combination with a COMT inhibitor found that self-reported FoG was alleviated in the group taking L-DOPS and entacapone, whereas this effect was absent the groups on L-DOPS or entacapone alone [106].

Another norepinephrine agent, droxidopa, has been reported to improve neurogenic orthostatic hypotension in PD [107]. Recent evidence suggests that normalizing orthostatic hypotension in PD subjects reduces the risk for falls [108]. No studies have evaluated the effects of norephinephrine agents on objective measures of gait.

Methylphenidate

The effect of methylphenidate is commonly used in attention-deficit disorders to enhance attention functions, acting primarily as a dopamine reuptake inhibitor [109]. In combination with levodopa, methylphenidate improved immediate levodopa effects on gait speed [110, 111] and temporal variability of gait [110] in people with PD. More recently, methylphenidate has been used in more advanced PD patients with treatment-resistant FoG after STN-DBS. A multi-center RCT reported improvements in FoG in these advanced patients [89], in line with previous non-RCT studies [110, 112]. However, using more specific measurements of gait in a small sample (n=12) in the same study did not provide evidence for improvement [113]. In subjects with PD with moderate gait impairments (no DBS), no beneficial effects of methylphenidate could be demonstrated, although the authors reported some slight improvements on methylphenidate in a gait composite score when off, but not when on dopaminergic treatment [114]. In summary, there is inconclusive evidence for a benefit of methylphenidate to improve gait function in PD.