Parkinson’s disease (PD) is a chronic neurodegenerative disorder, characterized by motor and non-motor symptoms. The typical PD clinical manifestations are motor control impairments such as tremor, muscular rigidity, and bradykinesia1. However, there is a wide host of non-motor neuropsychiatric impairments implicated in PD, such as anxiety, apathy, cognitive dysfunction, and depression. These neuropsychiatric symptoms are especially debilitating and affect PD patients’ quality of life (QOL), yet may be under-reported2. For example, there is an evidence that depressive symptoms impair QOL and functioning more than any other PD motor and non-motor symptom3. Depressive symptoms are reported as high as 89% in the PD population4, with a mean reported prevalence rate of 40% in outpatient and 54% in inpatient settings5. Other non-motor symptoms affect QOL at the early stages of PD. In an exploratory drug trial, the most frequent psychiatric symptoms in PD patients were irritability (66.1%), depression (48.3%) followed by apathy (40.3%)6. While meta-analyses estimated more modest rates of 39% for depression (17% for major depressive disorder and 22% for minor depression)5, 31% for anxiety7, and 39.8% for apathy8. Symptoms of PD depression (PD-dep) are clinically different than symptoms in general depression, and more often portray severe irritability, sadness, dysphoria, pessimism, and suicide ideation9. The etiology of PD-dep is thought to be particularly influenced by interactions between exogenous (i.e., diagnosis of a chronic and disabling disease) and endogenous causes (i.e., loss of dopamine)10. The clinical manifestations of PD are elicited by the progressive loss of dopamine neurons. Disruption of dopamine11,12 and glutamate neurotransmitter systems is implicated in the heightened vulnerability and loss of dopamine neurons. The involvement of the glutamatergic system in modulating psychiatric disorders was first proposed by altered glutamate receptor expression13 and altered glutamate–glutamine levels in cerebrospinal fluid of patients with mood disorders14.

Abnormal glutamate signaling

Alterations in glutamatergic transmission are implicated in PD pathophysiology. The most characterized receptor in glutamate neurotransmission is the N-methyl-D-aspartate (NMDA) receptor. The NMDA receptor is composed of heteromeric subunits (NR1 and NR2), a glycine binding site, and a glutamate binding site15 (Fig. 1). The activation of NMDA receptors requires co-agonist binding of glycine/D-serine and glutamate; therefore, antagonists that disrupt co-agonist binding, effectively block the NMDA activity. The hyper-phosphorylation and resulting overactivation of NMDA receptors is well-established in PD; and is implicated in the worsening of dyskinesias16,17,18. The short-term L-DOPA-induced dyskinesias (LIDs) are a debilitating side effect of L-DOPA administration, and NMDA receptors are presumed to be partially responsible for LIDs19. The LIDs are a severe therapy-related complication in PD, and significantly impair QOL. Positron emission tomography (PET) images have confirmed an enhanced NMDA receptor activity in specific motor cortical areas of the brain during LIDs in PD patients20.

Fig. 1: NMDA receptor consists of two heterodimers.
figure 1

Each heterodimer contains two extracellular subunits: NR1 and NR2. The NR1 subunit contains the glycine binding site, whereas the NR2 contains the glutamate binding site. Arrows show possible binding sites of uncompetitive/non-competitive antagonists (orange) and competitive antagonists (white)

The use of NMDA antagonists in PD is supported by three observations: (1) blockade of aberrant glutamate signaling in the subthalamic nucleus is crucial in the pathogenesis and motor PD symptoms, (2) subthreshold doses of NMDA antagonists synergize with Parkinsonian and dopaminergic agents21 by causing enhanced release and turnover of striatal dopamine21, and (3) PD models suggest that NMDA antagonism may protect nigral neurons21,22 (Fig. 2). It has been demonstrated that not only does NMDA antagonism improve PD symptoms, but may also be neuroprotective, preventing disease progression by inhibition of glutamatergic-mediated excitotoxicity23, and stimulating synaptogenesis/neurotrophic release24,25.

Fig. 2: The motor and non-motor symptoms in Parkinson’s disease are hypothesized to arise from similar mechanism(s) involving a loss of dopamine input, leading to Glu hyperactivity.
figure 2

NMDA antagonists block hyperactive Glu binding with NMDA receptors and exert ameliorating effects on mood and motor function. Glu glutamate, LIDs levadopa-induced dyskinesias, AMPA 2-amino-3-(5-methyl-3-oxo-1,2-oxazol-4-yl)propanoic acid, BDNF brain-derived neurotrophic factor

Drugs for neuropsychiatric PD symptoms

Co-morbid depressive symptoms in PD patients are detrimental to daily life activities and there are indications that depression exacerbates cognitive and motor impairments in PD26. The impact of depressive symptoms, and the relationship between mood circuits, cognitive circuits, and PD, makes a compelling case for initiating treatment. Unfortunately, the response rate to first-line antidepressant medications is low among the geriatric PD population when compared to placebo27. Although selective serotonin reuptake inhibitors (SSRIs) are considered first-line therapy, their clinical efficacy in PD-dep is inconclusive28. In addition, SSRIs may lead to a worsening of motor symptoms, due to antagonistic effects on dopamine29. The Movement Disorder Society Task Force concluded that there was a lack of efficacious therapies in treating anxiety, apathy, and impulse control symptoms in PD30. There is a critical need for alternative pharmacological methods to treat PD-dep, as mainstream antidepressant drugs are largely ineffective in this population27,31,32. The use of memantine for treating cognitive dysfunction and other psychiatric symptoms (anxiety and depression) in PD was suggested by two small clinical trials6,33. There are anecdotal reports of PD subjects claiming “better mood” or “improved sense of humor” after memantine treatment34. Given the dual antidepressant and anti-dyskinetic effects of ketamine and other NMDA antagonists for PD, a number of NMDA antagonists will be reviewed for treating depression and other neuropsychiatric symptoms, including memantine, amantadine, ketamine, dizocilpine, and d-cycloserine. Ketamine is an agent that has been re-purposed for treating treatment-resistant depression35. The promise of ketamine in treating PD-dep is of particular interest. A recent case report series demonstrated the reduction of LIDs in PD patients after intravenous (IV) ketamine administration and the significant reduction in pain and depressive symptoms36. Although NMDA antagonists are classically utilized to reduce LIDs and motor symptoms in PD, the objective is to review the modulatory effects on neuropsychiatric symptoms. Given the few number of studies focused on PD-dep, behavioral and all cognitive symptoms will be included.


A literature search was conducted using PubMed database. We used the following keywords: NMDA antagonist and Parkinson’s disease. We also searched the following databases up to March 1, 2018: Ovid MEDLINE, PsycINFO, CINAHL, Google Scholar, Cochrane Central Register of Controlled Trials, and Cochrane Database of Systematic Reviews. The initial search yielded 263 articles. The title and abstracts of the articles were then scanned for keywords, such as mood, psychiatric condition, depression, and cognition, yielding 40 full-text articles. The following criteria were used for inclusion: (1) articles in English or with an available published English translation, (2) publication in a peer-reviewed journal, (3) studies which quantitatively or qualitatively described the change in mood, QOL, sleep, cognition, or neuropsychiatric symptoms after NMDA antagonist therapy. Two authors (B.V. and W.O.) independently reviewed the articles identified from the search using the above selection criteria. After a review of full-text articles, 20 articles were excluded since they were not in English (n = 4), did not describe the change in neuropsychiatric symptoms before and after therapy (n = 10), or were literature reviews (n = 6). The two authors reached consensus on the 20 studies included.


Description of included studies

Of the 20 articles included in this review, there were eight randomized placebo-controlled trials, one double-blind placebo-controlled study, four open label studies, one washout placebo-controlled trial, two case reports, and three animal studies. The pre-clinical studies included rat and monkey PD animal models. The neuropsychiatric symptoms that were described include depressive symptoms, irritability, anxiety, sleep, QOL, global neuropsychiatric inventory scores, executive function, and changes in cognition. The findings of the studies are listed in Table 1. For the full descriptions of the acronyms of the various clinical scales used, refer to Table 2 and for descriptions of each drug, refer to Table 3.

Table 1 List of articles on NMDA antagonists and Parkinson’s disease in PubMed Database (1969–2018)
Table 2 Description of the scales and measures commonly used in studies
Table 3 Description of NMDA antagonists and mechanism of action(s)


There were 11 human studies that described the neuropsychiatric symptoms with memantine treatment. All trials were conducted at 20 mg/day, unless noted otherwise.

The effect of memantine on cognition was assessed in five randomized controlled trials (RCTs), and measured by the mini-mental state examination (MMSE), neuropsychiatric inventory (NPI), or clinical dementia rating (CDR) scales. In a 24-week trial, the memantine group had a slight, yet significant improvement in cognition on the MMSE (p = 0.02). Overall, there was a greater improvement in the memantine group as 27% of patients experienced a moderate to substantial improvement and no placebo patients reported more than a slight improvement. However, there was an insignificant change in NPI scores37. In a separate 24-week trial, 30 PD patients experienced improvements in cognition with medium to large effect sizes in information processing and recognition memory38. In a 24-week trial with memantine, the NPI scores did not significantly differ in PD patients (p = 0.522)39. Although there were notable improvements in items such as apathy, anxiety, irritability, and depression, the improvements did not reach statistical significance when compared to placebo. Cognitive scores were insignificant for PD (p = 0.576). A placebo-controlled 24-week trial of memantine improved cognition among 32 PD patients, as evidenced by various scales, including the Alzheimer’s disease assessment scale-cognitive (0.002), the frontal assessment battery (p = 0.01), verbal fluency test (p = 0.01), and clock drawing test (p = 0.03). Notably, the improvement in NPI sub-sections included disinhibition (p = 0.006), irritability (p = 0.004), anxiety (p = 0.04), and hallucinations (p = 0.048) when compared to placebo. The most significant change was caused by the decreased disinhibition/impulsive behavior in four patients. The number of patients who improved was not specified among the other NPI sub-sections. There was no significant change or improvement with memantine at 12 weeks6. When treatment was extended to 52 weeks, there were significant improvements and enhancement on cognition and neuropsychiatric symptoms on all scales (p < 0.05).

The cognitive effects of memantine discontinuation were shown in a 22-week trial by Leroi et al., 25 PD patients were entered in a washout phase after a 16-week trial with memantine. After the washout phase, patients were re-assessed at a 6-week follow-up. At follow-up, the improvement in cognition was only evident with the dementia rating scale (DRS), with insignificant differences between memantine and placebo with the NPI and MMSE scales, the sub-scores for each scale were not shown. At follow-up, the memantine group also had more global deterioration (70%) than placebo (29%)33. In a washout trial and post 24-week treatment with memantine, the attrition rate was most affected by the worsening of anxiety and depressive symptoms (9%). The clinical global impressions (CGI) scores also considerably worsened in the memantine washout group vs. placebo40. A case report (n = 2) described an improvement on mood/behavior for one patient after 3-month treatment, but no change in the other patient, measured by the unified Parkinson’s disease rating scale (UPDRS)-I section41.

The changes in QOL in PD patients after memantine treatment were insignificant in an 8-week trial34 and in a 22-week trial42, but significant improvements were observed in a 24-week trial in which the memantine group had a 42% higher QOL compared to 15% placebo (p = 0.01)43. In terms of sleep quality, an 8-week exploratory pilot trial by Ondo et al. did not show a significant change in sleep34, while the 24-week trial observed improved sleep and less physical activity during sleep compared to placebo (p = 0.006), without affecting daytime sleepiness44. The Ondo et al. trial recorded insignificant changes in other mood metrics, such as in UPDRS-I, II, the fatigue severity scale, and the Hamilton depression scale34.


There were four studies that described the neuropsychiatric symptoms after amantadine treatment. The safety and efficacy of extended release amantadine (260, 320, and 420 mg/day) were established in an 8-week randomized, double-blind controlled trial. Although the formulation was well-tolerated, there were insignificant changes in the global UPDRS and QOL scores at any dose. However, the mood score on the UPDRS-I section was unknown since it was reported as one global score, and incorporated the II–III sections45. The following three studies were open-label, non-randomized, and not placebo-controlled. Only one study had a cognitive measure, in which the visual and cognitive processing significantly improved after a 6-month trial of tapered amantadine (100–300 mg/day)46. During the course of a 6-week trial of tapered amantadine (100–500 mg/day), 26 of the 43 PD patients subjectively reported an “improved mood”, regardless of amantadine dose. There were no objective measures in the study47. Lastly, psychiatric adverse events were reported in a 6-month trial of amantadine (100–200 mg/day), as 20% of patients experienced increased jitteriness, insomnia, abdominal uneasiness, loss of appetite, and one patient developed depression48. These adverse symptoms promptly disappeared within 36 h of drug cessation. The authors commented that amantadine potentiated side effects of belladonna-like drugs, as those patients who reported adverse events were concurrently taking trihexyphenidyl (Artane) and benztropine (Cogentin)48.


Ketamine is an agent that has been re-purposed for treating mood disorders and effective at treating major depression49. Though there are a few studies discussing the use of ketamine for management of PD dyskinesias50,51. There was only one study to describe behavioral symptoms by researchers at Arizona Tucson University36. Five PD patients were treated with ketamine for a constellation of intractable pain and painful dyskinesia, and one patient had suicidal ideation and depression. Among both patients whose pain was assessed, there was a 50% decrease in pain after ketamine infusion (pain scale assessment, 1–10). The dose range was 0.05–0.15 mg/kg for 65 h or 96 h. Severe depressive symptoms and suicidality in one patient improved to “mild” depression after a 65 h continuous ketamine infusion at an average dose of 0.09 mg/kg/h. No metrics or scales were used to determine improvements in depression. This study suggests that low-dose sub-anesthetic ketamine infusions are well-tolerated and safe in a PD population.


There were two animal studies that measured the effect of d-cycloserine on behavioral and cognitive performance. In both studies, animals were challenged with chronic administration of 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridin (MPTP) to induce a PD model and subsequent administration of d-cycloserine to assess the extent of recovery in PD symptoms. The cognitive-enhancing ability of d-cycloserine for PD was suggested in a primate model after Schneider et al. treated primates with a single administration of d-cycloserine (320 or 1000 µg/kg), and there was a significant improvement in a variable-delayed-response task. However, there was no effect at a high dose of d-cycloserine (8000 µg/kg)52. In a PD rat model, Ho et al. observed improved cognition and memory after intraperitoneal injection of d-cycloserine across 13 days. In addition to memory, anxiety-like behavior decreased, and was observed by a number of behavioral animal tests, including the rotarod test, T-maze, and plus-maze53.


In a primate PD model, a single administration of dizocilpine (10–32 µg/kg) had no effect on cognitive improvement52. A PD rat model was challenged with the neurotoxin 6-hydroxydopamine (6-OHDA), and produced anxiety and depressive-like behavior. After intraperitoneal dizocilpine (0.2 mg/kg), rats spent more time in the light chamber (p < 0.01), and more contact time during the social interaction test (p < 0.05) in comparison to 6-OHDA lesioned rats. This suggests that MK-801 ameliorates depressive-like behavior in 6-OHDA lesioned rats54.


In an open-label pilot study of 10 PD patients, a daily dose of 90–180 mg/day of dextromethorphan did not significantly improve UPDRS sub-scores in mentation, behavior, and mood domains. The follow-up was at 1 month55.


Cognitive dysfunction

Neurocognitive dysfunction and dementia can occur in up to 40% of PD patients, with dementia onset being more associated with disease progression56. Similar to Alzheimer’s dementia, it is suggested that impaired cholinergic pathways are the cause of PD dementia57, however, this may be related to aberrant glutamate activity58. In addition to improvements in motor symptoms, there is preliminary evidence that NMDA antagonists may be effective at treating PD psychiatric symptoms. In most studies, NMDA antagonists significantly improved cognition as the primary outcome (n = 4)6,37,38,46. Drug discontinuation and washout with memantine caused significant global deterioration in cognition and overall neuropsychiatric symptoms, which may be indicative of the effectiveness during the dosing period (n = 2)33,40. The exceptions were two memantine studies that showed moderate improvement in cognition, but were not statistically significant when compared to placebo33,39. The utility of memantine may be specific for cognition and plays a minimal role for other psychiatric symptoms. For example, a single study showed significant improvements with the MMSE measure, but not with NPI or UPDRS, measuring changes in neuropsychiatric and global PD symptoms, respectively37. A systematic review and analysis compared memantine to cholinesterase inhibitors in treating PD dementia and found that both drugs significantly improved global impression; however, cholinesterase inhibitors were more effective at enhancing cognitive function than memantine59. Although less commonly used in PD dementia, amantadine showed promise in enhancing cognition, especially in improving the visual-cognitive processing and visual discrimination46. The therapeutic effect of longitudinal use of amantadine was suggested in an 8-year study, revealing that amantadine delayed the onset of PD dementia by ~ 3.2 years and attenuated dementia severity in a dose-response manner60. While other NMDA antagonists, such as d-cycloserine improved cognitive scores, the disadvantage is it has only been shown in PD animal models52,53. In summary, there is potential for NMDA antagonists not only in treating LIDs, but also in attenuating dementia.

Depressive and anxiety symptoms

For many of the included studies, changes or improvements in mood were qualitative. However, when measures were used, the NPI or the UPDRS-I was utilized to measure depression, anxiety symptoms, and “general behavior mentation and mood”. The NPI contains a wide inventory of psychiatric sub-score symptoms, such as anxiety, depression, and irritability. The use of amantadine for treating neuropsychiatric symptoms was first suggested in 1968, when it was noted that the drug appeared to produce a positive effect and a feeling of general well-being and affect among PD patients48 and 2 years later, self-reports of “improved mood” emerged among a PD cohort47. Memantine appeared to have improved anxiety and irritability symptoms to a greater extent, than other psychiatric symptoms described in the NPI6. One study described the anecdotal reports of PD patients endorsing a “better mood” with memantine41. Three memantine studies offset these observations by not showing a significant difference in global NPI scores between treatment and placebo37,39. The disadvantage was that only the global NPI scores were reported; there could have been significant changes in sub-scores, yet not in the global score. It is of clinical interest to report changes in these domains that are indicative for specific symptoms (anxiety and depression) between studies. Future studies would be strengthened by reporting of NPI sub-scores/domains. However, measuring depressive symptoms utilizing a semi-structured interview, such as the HAM-D scale, did not show an improvement. However, this trial was only 8-weeks long with memantine34, as opposed to the more common period of 24 weeks. The beneficial effect on mood may take longer, as a similar study showed that NPI scores improved at 24 weeks, but not at 12 weeks6. Memantine has shown previous success at treating obsessive-compulsive syndromes61,62. Notably, memantine has markedly decreased impulsive behavior in PD patients6,63, which may be explained in terms of the glutamatergic dysfunctions in the lateral orbitofrontal circuit. This same study showed a significant improvement in anxiety at 52 weeks, but not at 12 weeks. Washout of memantine can cause worsening of anxiety and depressive symptoms40. These detrimental effects observed during washout may also be indirectly indicative of memantine’s effectiveness.

The efficacy of ketamine as an antidepressant and its application for major depressive disorder is under active investigation. To date, there have been nine high-quality RCTs, which have documented the markedly high response rate in the ketamine intervention group, when compared to placebo49. Notwithstanding some limitations, ketamine is a promising therapy for treatment-resistant depression35,64 and may resolve suicidal thoughts in patients experiencing suicidal ideation after a single infusion65.

A continuous ketamine infusion resolved suicidal thoughts and severe depression in a PD patient. The patient’s dyskinesia and pain symptoms were also resolved after ketamine36. This highlights the efficacy of a single drug to treat a constellation of PD symptoms and the ability to improve QOL. Other researchers have started similar initiatives66. Given the aberrant NMDA signaling in PD pathology, the use of NMDA antagonists such as ketamine may be a viable therapeutic option. Although the initial results are promising, large-scale clinical trials will be needed to determine the efficacy and safety in PD64. The efficacy of ketamine has spurred the development of alternate glutamate modulating antidepressants, such as rapastinel67. An advantage of newer therapeutics such as rapastinel is the induction of antidepressant effects without the negative psychoactive side effects of ketamine68. To note, although there are antidepressant effects of other NMDA antagonists such as dextromethorphan69, there is no literature of their effectiveness in PD-dep. There are reports of neuroprotection after acute dextromethorphan in a number of PD models52,70. A recent study showed improvement in anxiety and depressive symptoms in a PD rat model with dizocilpine treatment, which was speculated to have occurred via Wnt signaling54. In summary, the use of NMDA antagonists is becoming more common in treating psychiatric symptoms, especially depression. It is likely that future therapeutics will become more focused on modulating the glutamate system.

Overall quality of life

The non-motor PD symptoms are major predictors in the decline of QOL. There are few studies that measured QOL after NMDA antagonist therapy. One study showed a 42% higher QOL after 24-week memantine therapy than at baseline, compared to 15% of placebo43. Although promising, a significant QOL improvement was not attained in two other memantine studies34,71 and one amantadine study45. However, the amantadine study duration was only 8 weeks and it is suggested that improvements in QOL may not be apparent until longer follow-up. For example, the change in QOL was insignificant with memantine at 8-week follow-up34,45, but significantly improved at 24 weeks43. There is an evidence that memantine may not only improve QOL, but may also have a disease-modifying effect. A longitudinal study of 227 newly referred PD patients found that non-motor symptoms’ burden and QOL were predictive in PD progression over a 2-year period. Specifically, sleep/fatigue, mood/apathy, and attention/memory domains were most significantly predictive of QOL changes72. Similar studies were conducted in PD patients in Northeastern Mexico and Taiwan. Specifically, sleep/fatigue, mood/cognition, and gastrointestinal domains were associated with worse QOL in the Mexican PD population73. For Taiwanese PD patients, the depression/anxiety item was strongest, where disease duration and severity, but not pharmacological therapy, were major predictors of non-motor symptoms74. Further development and evaluation of interventions that improve QOL are needed. Long-term follow-up studies show that the survival time was significantly higher in PD patients who were on memantine trials (p = 0.045)75. It is suggested that an early positive response to memantine may be a prediction factor in longer survival. Therefore, the clinical use of memantine and amantadine for improving QOL may be beneficial in disease progression.

Other considerations

The positive effects of memantine in PD symptoms may be explained by indirect effects of NMDA antagonism. It is suggested that NMDA antagonists may offer neuroprotection by counteracting the hyperactive metabolism in PD basal ganglia. In a cohort study, PET scans of PD patients on 6-week memantine treatment showed neuronal modulation in the basal ganglia with decreased regional cerebral blood flow in the basal ganglia76 and by blocking excitotoxicity77. The use of NMDA antagonists as mainstream medications for PD treatment is supported by various factors: (1) improving LIDs by pharmacological synergism with dopaminergic agents (i.e., L-DOPA), thereby, potentially decreasing the effective dose of L-DOPA21,78, (2) improving basal motor function by endogenous release of striatal dopamine in vivo79,80, and (3) offering possible neuroprotection in the context of overactive glutamatergic neurotransmission21,22 and possible alleviation of psychiatric symptoms, discussed in this review. However, widespread use is limited by intolerable side effects and the pharmacology of current NDMA antagonists. For example, amantadine affects other neurotransmitter systems that need to be carefully monitored in PD patients, such as norepinephrine, serotonin, gamma-aminobutyric acid and acetylcholine81, while memantine (Namenda) is well-tolerated and has few side effects82. Future research will need to focus on specific antagonists that act on aberrant NMDA antagonism.


The future of antidepressants will extend beyond modulating the serotonergic and dopaminergic neurotransmitter systems. The repurposing of existing NMDA antagonists, such as ketamine, for depression and newer therapies, such as rapastinel, highlight the movement toward modulating the glutamatergic system. There is preliminary evidence that NMDA antagonists may modulate psychiatric symptoms in PD. However, the current evidence is inconclusive, and further trials must be conducted to elucidate their psychiatric-modifying effects.

Strengths and limitations

Given the novel application of NMDA antagonists for treating non-motor PD symptoms, a broad approach was taken in reviewing the literature. This allowed for the discussion of preclinical models and self-reported or anecdotal symptom reports by patients and clinicians. One common limitation in the included trials was the use of concurrent medications. Many studies included patients concurrently taking L-DOPA and other dopaminergic-modulating PD drugs. This may confound the therapeutic effect of NMDA antagonists on mood. However, this may enhance the applicability to naturalistic settings. Another confounding factor was the heterogeneity of the study population. Especially relevant in the cognitive studies, patients with PD and Lewy body dementia (LBD) were treated and analyzed simultaneously. Notably, memantine appeared to fare better among LBD patients than in PD, as evidenced by greater improvements in CGI and NPI scores. However, there are reports of worsening of psychotic symptoms with memantine in advanced LBD, though it may be exacerbated due to multiple psychotropic medications83. These findings suggest that LBD and PD-dep are sufficiently different to warrant different dosing strategies with memantine. When possible, results from only the PD cohort were reported in this review.