Pain in the Neurodegenerating Brain

Insights Into Pharmacotherapy for Alzheimer Disease and Parkinson Disease

Timothy Lawn; Yahyah Aman; Katarina Rukavina; George Sideris-Lampretsas; Matthew Howard; Clive Ballard; Kallol Ray Chaudhuri; Marzia Malcangio


Pain. 2021;162(4):999-1006. 

In This Article

Parkinson Disease

Pain is a prevalent nonmotor symptom in people with PD (PwP), acknowledged by James Parkinson in 1817,[112] affecting 68% to 85% of patients.[13,23,103,116,127] Despite this, it remains underdiagnosed and undertreated.[6,31,41,58,83,156] Pain in PwP is multifaceted and may result from comorbidities, be caused or amplified by motor symptoms, and is subject to abnormal nociceptive processing, as PD-specific neurodegeneration affects peripheral, spinal, and cerebral pain pathways.[42,125] Attempts have been made to synthesize a clear picture of heterogeneous pain in PD (Table 1);[6,58,151] however, to date, our basic understanding of the relationship between PD pathophysiology and pain remains underdeveloped. Identifying well-defined subtypes, and elucidating their concomitant underlying mechanisms, should facilitate the development of personalised treatment of pain in PwP.[24,143]

Pain Processing is Altered in Parkinson Disease

Studies have largely reported reduced pain thresholds (greater sensitivity to pain) and lower pain tolerance in PwP (for meta-analysis, see Ref.[141]). Interestingly, no relationship between pain sensitivity and disease duration was reported across 26 studies.[141] Moreover, significant heterogeneity is seen within and across studies suggesting considerable interindividual differences with multiple contributory factors. Surveys have found intensity and frequency of pain to be higher in patients with more advanced PD; however, this likely reflects an increased incidence of musculoskeletal pain.[141] A study using quantitative sensory testing failed to find a difference between drug-naive pain-free patients and controls suggesting that abnormalities may arise later in the disease duration, relate to dopaminergic therapy, or be associated with the development of chronic pain.[62] In the absence of longitudinal investigation, the effects of disease progression are impossible to delineate but the power advantages of meta-analysis add credence to the possibility that enhanced pain sensitivity is engaged at a certain point during pathogenesis with a strong ceiling effect. Early pathophysiology within the midbrain and brainstem regions may therefore be important for elevated psychophysical pain sensitivity and reduced pain thresholds. Conversely, conditioned pain modulation paradigms, which assess the functionality of descending modulatory mechanisms, have been found to be comparable in controls and patients with PD in both ON and OFF states.[68,69] However, trend significant differences were seen between PD subtypes (akinetic rigid, tremor dominant, and mixed). Given the low power of the study, this supports the heterogeneity of pain processing in PwP and emphasises the need for large studies that allow for adequately powered substratification.

Functional magnetic resonance imaging has revealed maladaptation of pain networks present even at early disease stages in pain-free PwP compared with healthy controls. Increased pain-related BOLD activation was observed in the somatosensory cortex, cerebellum, and caudal pons.[138] Furthermore, activity in descending pain modulatory regions, such as the dlPFC, dorsal ACC, and subgenual ACC, is lower in PwP than in healthy individuals, and connectivity between dorsal ACC and dlPFC during anticipation of pain is reduced.[138] The bilateral activation of the nucleus accumbens (NA) in PwP is also lower than that in healthy controls, suggesting altered processing of cognitive and evaluative facets of pain.[120,140] A network-based analysis has shown dysfunction in reward pathways in PwP suffering from persistent pain, but not those without, with disconnection of the right NA and left hippocampus.[118] The NA has been implicated in the transition from acute to chronic pain across a variety of human and animal studies.[8,29,51,56,155] The direction of causality remains unclear, but dysfunction of reward and modulatory networks may predispose PwP to develop chronic pain and offer therapeutic targets.

Pharmacotherapy of Pain in Parkinson Disease

Pain in PwP remains neglected and poorly understood, with only a minority of patients receiving adequate treatment.[13] People with PD are more likely to be prescribed analgesics, such as opiates, acetaminophen, antiepileptics, and antidepressants, as well as receive chronic prescriptions, risking polypharmacy or burdensome side effects.[22] Dopaminergic replacement therapy might lead to pain relief in some PwP.[92,142] For example, a 2-fold improvement in the King's Parkinson Disease Pain Scale domain "fluctuation-related pain" was observed with rotigotine vs placebo.[124] L-Dopa administration reversed the reduction of pain threshold seen in PwP during the off-state[64] and normalised abnormally increased pain-related activation within sensory-discriminative (insula) and cognitive-affective (prefrontal cortex and ACC) regions in a positron emission tomography study.[21] Interestingly, pain reduction from L-dopa administration or deep brain stimulation [for review, see; Refs.[39,45,91]] does not correlate with motor improvement suggesting it may act directly on pain circuitry.[40,92,102,142]L-Dopa is not only converted exclusively into dopamine but also into noradrenaline and may act as a false neurotransmitter within serotonergic terminals.[50] As both monoamines play a role in descending pain modulation and are affected by PD-specific neurodegenerative changes at prodromal stages, the pain modifying effect of L-dopa may be partially mediated through nondopaminergic systems.[9,19,20,44,74] Accordingly, duloxetine led to some degree of pain relief in an open-label study.[49] Cannabis has shown an ability to markedly reduce both sensory and affective facets of pain in PwP.[132] Interestingly, an oxycodone RCT failed to reach significance for the primary end point of reducing 24 hour pain scores.[144] There was a trend reduction in pain, and the dosage may have been inadequate. However, opioidergic circuitry is known to be perturbed by PD pathophysiology, and this may affect the efficacy of opioid analgesia.[54,115,136,140] Safinamide, with actions on dopamine through monoamine oxidase-B inhibition as well as modulating abnormal glutamate release, has also shown a benefit in PwP.[26,27,65] Rotigotine, a purely dopaminergic agonist, produces limited benefit for overall pain in PwP suggesting that safinamide may well impart a benefit through glutamatergic actions and this warrants future investigation.[124] However, there remains a paucity of robust studies with the Movement Disorder Society non-motor symptoms treatment recommendation identifying only 2 as sufficiently high quality to include.[131] The multiplicity of neurotransmitter systems through which these drugs act eludes to the complexity of pain in PD. Future research should use refined populations, or those large enough for substratification, to further elucidate how these interventions differentially interact with PD subtypes.

Utility of Animal Models

Animal models offer a unique opportunity to probe mechanisms of pain and pharmacotherapy. This has been well reviewed for PD,[55,147] but remains understudied in AD. Mirroring clinical populations most studies report altered pain thresholds compared with controls.[7,59,67,94,99,105,133,137] A chemically induced model of osteoarthritis through an intra-articular injection of monosodium iodoacetate within transgenic TASTPM AD mice has provided insights into interactions between clinically relevant pain, neurodegenerative pathophysiology, and opioid analgesia.[4,5] TASTPM mice demonstrate an age-dependent reduction in thermal nociception that coincides with amyloid pathology in pain-related brain regions.[4] Naloxone, an opioid antagonist, restored thermal nociceptive thresholds to that of wild-type controls. Mice modelling with combined AD and osteoarthritis exhibited impaired mechanical hypersensitivity and a lack of weight asymmetry. Subsequent administration of morphine not only produced an antinociceptive effect but also increased the noxious threshold significantly greater than that seen in wild-type animals.[5] Conversely, gabapentin showed no efficacy. Thus, altered processing within opioidergic circuitry may partially mediate altered pain processing as well as influence both efficacy and centrally mediated side effects of opioidergic pharmacotherapy. Additional preclinical investigation may yield similar avenues for translational investigation.