How to Use the New European Academy of Neurology/Movement Disorder Society European Section Guideline for Invasive Therapies in Parkinson's Disease

Abstract

Background

The decision to choose invasive treatments for Parkinson's disease (PD) is complex and needs careful consideration.

Objectives

Although the recommendations of the European Academy of Neurology/Movement Disorder Society European Section guideline for invasive therapies of PD are useful, the different clinical profiles of people with PD who seek advice for possible invasive therapy need further attention.

Methods and Results

Here we describe 8 clinical standard situations of people with PD unsatisfied with their current oral treatment where invasive therapies may be considered. These are PD patients presenting with the following symptoms: (1) severe motor fluctuations, (2) beginning of levodopa‐responsive fluctuations, severe tremor at (3) young or (4) advanced age, (5) impulse control disorders and related behavioral disorders, (6) hallucinations and psychosis, (7) minimal cognitive impairment or mild dementia, and (8) patients in need of palliative care. For some of these conditions, evidence at lower level or simple clinical considerations exist.

Conclusions

There are no one‐fits‐all answers, but physician and patient should discuss each option carefully considering symptom profile, psychosocial context, availability of therapy alternatives, and many other factors. The current paper outlines our proposed approach to these circumstances.

Management of Parkinson's disease (PD) comprises a range of different therapy approaches, some being favored at certain disease stages. ¹ In the course of PD progression, oral therapy often cannot provide sufficient symptom control, and complex combined treatment involving nonmedical, medical, and invasive therapies is required. ² Thus, the identification of advanced PD patients suitable for invasive treatment options is crucial for promptly offering personalized therapy and, thus, improving the quality of life of PD patients. ³

The development of invasive treatments for PD has a long history, but the number of options has never been as rich as nowadays. Mostly, these interventions are indicated and managed by specialized centers. However, it is important also for the general neurologist to consider these treatments as an option, and often the treating physician's choice is the door opener for such treatments. ⁴ , ⁵ Therefore, the European Academy of Neurology (EAN) and the Movement Disorder Society European Section (MDS‐ES) have recently issued a guideline (GL) on invasive therapies for PD ⁶ , ⁷ replacing in part former EFNS/MDS‐ES GLs ⁸ , ⁹ and MDS‐EBM reviews. ¹⁰ , ¹¹ This new EAN/MDS‐ES GL is detailed, but its usage to counseling specific PD patients may be difficult for neurologists not specialized in movement disorders. Thus, we have chosen different typical clinical scenarios to discuss some considerations for or against the choice of such interventions based on the GL, literature that does not fulfill inclusion criteria of the GL and clinical considerations. Although the basis for treatment selection is the evidence and patients' preferences, costs of care and availability of invasive therapies are also important for the decision‐making process. This paper aims to examine specific clinical scenarios when invasive therapies should be considered.

Patients and Methods

The authors collected data on their inpatients and outpatients over a 1‐year period requesting advice for invasive therapies. The lead authors (D.B. and G.D.) classified these approximately 300 PD patients for their main clinical problem. This led to overlapping classes that were broken down by the lead authors into 8 scenarios of main clinical problems. This was presented to the whole author group, discussed, and finally approved by all authors. These are 8 types of patients with PD presenting with the following symptoms: (1) severe motor fluctuations, (2) beginning of levodopa (l‐dopa)‐responsive fluctuations, severe tremor at (3) young or (4) advanced age, (5) impulse control disorders (ICD), (6) hallucinations and psychosis, (7) minimal cognitive impairment or mild dementia, and (8) patients in need of palliative care. PubMed (2010–2022) was searched using key words “DBS,” “apomorphine,” “LCIG,” “focused ultrasound,” and each of the 8 patient types. The authors formally agreed on the different options to be discussed for each type.

The author group has had access to the original data and forest plots of the EAN/MDS‐ES GL. From the cumulated results of relevant outcome parameters based on the Forest plots of the GL (see Appendix 4 of the GL ⁵ ), the difference in mean change between the interventional group and the medically treated control group of the same study/meta‐analysis was calculated. These results are shown for relevant parameters in Figure 1.

FIG. 1.

Effect of the different interventions displayed as the difference in mean change (±95% confidence interval) between the interventional group and its study‐related control group for each of the interventions with control groups (original data, see Appendix 4 of the new guideline). Clinical head‐to‐head studies comparing the interventions are not available. In the case of severe adverse events the risk ratio and in the case of daily dose of levodopa (l‐dopa) equivalent the mean reduction (in mg) between the interventional and control groups are shown (Apo, apomorphine pump; DBS, deep‐brain stimulation; EarlyStim, deep‐brain stimulation in patients with PD with beginning fluctuations; FUS, focused ultrasound; LCIG, levodopa‐carbidopa intestinal gel; STN, subthalamic nucleus) cumulated from the RCTs (randomized controlled trial). The dot size indicates the number of subjects included in the trials. UPDRS (Unified Parkinson's Disease Rating Scale) II evaluates the motor aspects of experiencing daily living, UPDRS III rates the motor assessment in medication off (except for Apo and LCIG), and UPDRS IV evaluates the motor complications. Life quality estimated using the Parkinson's disease Quality of life Scale‐39 (PDQ‐39) except for Apo where the PDQ‐8 summary index was used. On‐time was defined as on‐time without troublesome dyskinesia (hours per day). Serious adverse events are defined as events that indicate further interventions or hospitalization. Daily dose of l‐dopa equivalent was calculated in each publication based on accepted conversion methods showing the daily intake of antiparkinsonian medication.

Results

Current Overall Recommendation for PD Patients

According to the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) ¹² methodology, the GL group included only randomized controlled trials (RCT) in the GL. ⁷ For the strength of the available evidence, a clear ranking for the different interventions is given. The strongest evidence is for deep‐brain stimulation of the subthalamic nucleus (STN‐DBS, 5 RCTs ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ ) followed by DBS of the internal globus pallidus (GPi‐DBS) against STN‐DBS (2 RCTs ²⁰ , ²¹ ) and l‐dopa–carbidopa intestinal gel (LCIG) pumps (2 RCTs ²² , ²³ ). Lower evidence levels are found with respect to continuous apomorphine injection (CAI) pumps (1 RCT ²⁴ ), unilateral lesioning of the STN with magnetic resonance imaging‐guided focused ultrasound lesioning (MRgFUS subthalamotomy, 1 RCT ²⁵ ), and pallidotomy with radiofrequency thermocoagulation (2 unblinded RCTs ²⁶ , ²⁷ ). The evidence for other lesional procedures (radiofrequency, radiosurgery) or MRgFUS in other targets than STN is low; therefore, recommendations were not possible, and clinical consensus statements are proposed.

The interventions have been tested in the pivotal studies on patient groups with well‐defined profiles. Certainly, many patients will not exactly fit these descriptions, and further arguments are needed for or against the application of the treatments for these individuals.

One argument is always the efficacy of a treatment compared to best medical treatment. Figure 1 shows the difference in mean change between the interventional group and the corresponding control group with standard treatment as taken from the Forest plots of the GL. ⁷ Whenever the 95% confidence limits cross the zero line, the difference between the intervention and the medical control groups is not significant. This measure does not compare different interventions but gives an estimate of the effect size of the outcome parameters for the different interventions. We note that since the publication of the GL a randomized, sham‐controlled MRgFUS trial was published and is now FDA approved. ²⁸ Similarly, subcutaneous infusion of l‐dopa preparations and the addition of entacapone to l‐dopa/carbidopa infusions have been approved in Europe. These interventions have not been evaluated by the GL committee. This paper will not prejudge the pending statement of the GL commission.

Clinical Decision‐Making for Specific PD Symptom Profiles

Attempts have been made to identify patients who need invasive therapies using consensus criteria. ⁵ The simple “5‐2‐1” rule (oral l‐dopa intake 5 times a day, OFF symptoms at least 2 hours a day, and troublesome dyskinesia for at least 1 hour a day) ³ , ²⁹ was found to be a reasonable predictor for candidates for invasive therapies, notably patients with advanced PD.

We have grouped our patients based on their symptom profile. Considerations for these different patient profiles are discussed in the following sections of this clinical pathway. We do not intend to provide cooking recipes but to guide clinical thinking for such patients.

The Severely Fluctuating Patient

Clinical Profile and Therapeutic Aim

This is, by far, the largest group of patients seeking improvement in their clinical conditions with invasive therapies. Oral l‐dopa treatment in PD is associated with development of complications of the therapy, including motor and nonmotor fluctuations and drug‐induced dyskinesias. ²⁹ Most patients experience those complications after more than 5 years of therapy, and up to 30% develop these problems mildly within the first 2 years. ³⁰ Invasive therapies are considered for those patients (advanced stage of PD) when optimized l‐dopa therapy and adjunct therapies are no longer sufficient to increase on‐time and reduce off‐time according to the mutual agreement between the patient and physician.

GL, Lower‐Quality Evidence, and Long‐Term Observations

The most extensively studied intervention for advanced PD with medically unresponsive fluctuations is STN‐DBS surgery. In total 5 RCTs have shown that it results in a large improvement in motor impairment, motor fluctuations, and dyskinesia and with a large increase in daily on‐time in these patients. ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ A number of uncontrolled long‐term studies of patients with advanced PD and STN stimulation are available and underwent multiple metanalyses. ³¹ , ³² , ³³ These reveal that STN‐DBS is still effective beyond 15 years after the intervention. Also, GPi stimulation can be recommended equally to STN stimulation as comparative RCTs between the effects on the 2 targets showed only minor differences during 3 years of follow‐up. ²¹ , ³⁴ Concerning long‐term observations, the evidence is scarcer, but at least a retrospective GPi study with the longest follow‐up of 8 years showed a sustained improvement in dyskinesia and fluctuations although not reaching significance level when compared to baseline. ³⁵ Importantly, l‐dopa dosage cannot or can only slightly be reduced using GPi‐DBS in contrast to STN‐DBS. The decision on which of the 2 targets is best for a patient should be left to the treatment center.

Additionally, LCIG or CAI may be considered for this PD group. Two RCTs are identified on the LCIG treatment of advanced PD: one 3‐month double‐blind, double‐dummy study of LCIG against oral l‐dopa plus otherwise‐stable antiparkinsonian treatment and the DYSCOVER trial against optimized oral treatment. ²² , ²³ In general, on‐time without troublesome dyskinesia was probably moderately increased in the overall group (as shown in Fig. 1). There was a moderate decrease in daily off‐time in the LCIG group compared to oral treatment within the observation period of 3 months. Open‐label longer‐term (2 and 3 years) changes have been assessed within a registry and document a sustained reduction in off‐time and on‐time dyskinesia as well as an improvement in quality of life. ³⁶ , ³⁷ Despite the fact that the literature shows a sustained effect at least for up to 12 months, longer‐period controlled data are still needed. ³⁸

One RCT on CAI with a 12‐week double‐blind phase ²⁴ and a 52‐week open‐label phase ²⁴ , ³⁹ showed a moderate improvement in daily on‐time without troublesome dyskinesia in the apomorphine group compared to the Best Medical Treatment (BMT) group receiving placebo infusion. Similarly, there was a moderate improvement in daily off‐time in the apomorphine group against BMT. Quality of life and motor disability did not significantly improve. Multiple open‐label studies confirm the efficacy of CAI in the reduction of daily off‐time. ⁴⁰ , ⁴¹ , ⁴² , ⁴³ Despite several retrospective large case series, detailed long‐term effects are understudied and have been reviewed recently ⁴³ ; however, retrospective data confirm a sustained reduction in motor fluctuations for at least 2 years. ⁴⁴

Although effectiveness and safety of unilateral radiofrequency pallidotomy could be shown in EBM reviews for fluctuating PD patients, the GL recommends this only if other treatment options are not available. ⁷

Potential Limitations

The main limitations of these therapies are adverse events. Adverse events may involve surgery‐ or device‐related complications and neuropsychiatric symptoms, but also gastrointestinal complications in case of LCIG, dysarthria or gait impairment in case of DBS, and even dyskinesia in case of CAI. ¹³ , ¹⁶ , ²² , ²⁴ , ²⁵

Another limitation are contraindications. For STN‐DBS this comprises neurosurgical contraindications, neuropsychiatric disorders, and age limit, although the latter has not been scientifically explored so far. For LCIG, exclusion criteria include any condition interfering placement of a Percutaneous endoscopic gastrostomy tube or gastrointestinal absorption of the gel and severe cognitive or neuropsychiatric symptoms, although clear‐cut selection criteria are still lacking. ²² , ⁴⁵ Relative contraindications for CAI comprise mainly internal diseases (electrocardiogram alterations, hemolytic anemia, and local skin infections), severe psychiatric and cognitive symptoms (dementia), and orthostastic hypotension. ² , ²⁴ , ⁴⁶ For all interventions a lack of l‐dopa sensitivity at least at high dosages is a contraindication.

Clinical Reasoning

For this patient group, effective treatments are available. The criteria to choose a specific intervention or not include criteria like operability, availability of interventions, and probable adverse events. In particular, age and multimorbidity can limit the operability and the outcome of DBS. ⁴⁷ , ⁴⁸ , ⁴⁹ As cognitive impairment and mild psychiatric symptoms constitute relative contraindications for all invasive treatments, these specific PD groups are discussed separately. Apart from DBS, there is a lack of long‐term data; therefore, stimulation has the strongest evidence. For countries that have no DBS and pump therapies, unilateral pallidotomy may be an option.

PD with Early Fluctuations

Clinical Profile and Therapeutic Aim

The patients (defined as below 60 years in the EarlyStim trial ¹⁸ ) include those having motor signs for at least 4 years, those with excellent l‐dopa sensitivity (≥50%), and those experiencing at least mild fluctuations. ⁷ The main therapeutic goal is to optimize the quality of life during this period when patients still have a good response to dopaminergic medication but have a few symptoms that are difficult to control.

GL, Lower‐Quality Evidence, and Long‐Term Observations

Two RCTs examined people with PD aged below 60 years and with early fluctuations. ¹⁷ , ¹⁸ STN‐DBS resulted in a large improvement in quality of life and motor symptoms. Motor fluctuations and dyskinesia improved to a similar degree as in advanced PD. Gait freezing and neuropsychiatric fluctuations improved. Long‐term data are not yet available for this cohort.

The main study and secondary analyses suggested that important requirements for this indication are a 50% improvement in preoperative l‐dopa test and normal psychiatric and cognitive status. The quality of life before intervention should be relevantly reduced. ⁵⁰ Presence of fluctuations is particularly important, as 2 small studies on STN‐DBS in young patients without fluctuations did not report significant positive effects on symptoms and course of disease ⁵¹ , ⁵² ; therefore, the EAN/MDS‐ES GL does not recommend STN‐DBS for nonfluctuating patients. The extent of these fluctuations must not be severe and may consist of morning‐OFFs only in some patients, but it should be well documented according to the EAN/MDS‐GL. ⁷

For other interventional treatments, no higher evidence is available for this particular patient population. However, a nonrandomized prospective study on CAI in initial‐stage PD patients reported a benefit on motor symptoms, and the Early‐PUMP RCT is ongoing. ⁵³

Potential Limitations

The main limitation of STN‐DBS in young PD patients is the occurrence of adverse events, which are more common than in patients receiving BMT (54.8% vs. 44.1%). Adverse events may involve surgery‐ or device‐related complications, which have been resolved without sequela within 24 months in the EarlyStim study. ¹⁸ , ⁵⁰ Recently, it has been shown that the predictive value for the 1‐year outcome is less precise than previously thought, ⁵⁴ but this criterion will definitely avoid including nonresponders and has proven to be successful for EarlyStim.

Clinical Reasoning

STN‐DBS may be considered in PD patients with early fluctuations who fulfill inclusion criteria for early stimulation as it improves motor symptoms and quality of life relevantly better than medical treatment. All other interventions have not been tested in this patient group.

The Young Tremor Patient

Tremor is the most common initial symptom of PD across all age groups. ⁵⁵ The age limit for young PD patients was set at 60 years as in the EarlyStim cohort. First‐line therapy is dopaminergic agents although tremor shows an inconsistent response to oral medication. ⁵⁶

Clinical Profile and Therapeutic Aim

Young PD tremor patients seek advanced treatment options when oral medication does not sufficiently suppress tremor. ⁵⁷ Further, due to the progressive character of PD, symptoms are expected to become more numerous and severe in the future. Therefore, the aim of invasive therapy in this patient cohort is the improvement in present motor symptoms as well as a far‐sighted therapeutic option for prospective symptoms.

GL, Lower‐Quality Evidence, and Long‐Term Observations

DBS of STN, GPi, and Ventro‐intermediate nucleus of the thalamus (VIM) is known to effectively reduce tremor severity. ¹⁹ , ³² , ³³ However, only stimulation targeting STN and GPi influences other motor symptoms of PD. Therefore, these targets are the best choice for PD patients with a similar effectiveness. ⁵⁸ However, although l‐dopa responsiveness is a benchmark for DBS responsiveness, only STN‐DBS is reported to further control dopamine‐resistant tremor. ³³ , ⁵⁴

STN‐MRgFUS can be an alternative for tremor‐dominant PD patients with asymmetric tremor presentation as it achieves a tremor reduction up to 92% unilaterally. ²⁵ However, due to its novelty, evidence is based on only 1 RCT and lower‐quality evidence studies. ²⁵ , ⁵⁹ Long‐term data report a sustained benefit over 3 years for STN‐MRgFUS, ⁶⁰ whereas for DBS long‐term efficacy beyond 15 years is well established. ³² , ³³ Therefore, DBS should be preferred in younger patients, which allows adaptation to disease progression by increasing the stimulation parameters until long‐term data for STN‐MRgFUS are available. VIM‐MRgFUS has been reported in 1 underpowered controlled study ⁶¹ and several case series for parkinsonian tremor, ⁶² , ⁶³ , ⁶⁴ , ⁶⁵ but there is no sufficient effect on akinesia and rigidity to recommend this for patients with a life expectation of decades.

Although some literature studies report an antitremor effect of LCIG and CAI, there is limited evidence for these treatment options in orally insufficiently treated PD tremor. ⁶⁶ , ⁶⁷

Potential Limitations

Limitations of DBS are outlined in part 1. In the case of STN‐MRgFUS, possible adverse events and the limitation due to skull permeability for ultrasound represent treatment limitations. ²⁵ Patients are eligible only when symptoms are clearly asymmetric and if there is no history of brain hemorrhage, intracranial surgery, and current neuropsychiatric disorders. ²⁵ Adverse events are mostly transient procedure–related complications, for example, headache, nausea, or superficial wounds, and ablation‐related complications such as ataxia with gait disturbance, dysarthria, weakness, paresthesias, or sensory defects. ²⁵ , ⁵⁹ Finally, MRgFUS is recommended only within clinical registries and in experienced centers. ⁷

Clinical Reasoning

The best option for young PD patients with tremor poorly responsive to oral treatment is STN‐DBS or GPi‐DBS implantation due to the immediate and long‐lasting effect.

Elderly Tremor Patient

Overall, tremor is present in 79% of all PD patients, having a huge impact on daily living. ⁵⁷ The age limit for older PD patients was set to over 60 years.

Clinical Profile and Therapeutic Aim

Elderly patients with PD tremor seek advanced treatment options when suffering from drug‐responsive but insufficiently treated tremor or drug‐resistant tremor. ⁵⁷ Therefore, the therapeutic aim for this patient cohort may be tremor suppression only.

GL, Lower‐Quality Evidence, and Long‐Term Observations

DBS is the only advanced therapeutic option controlling tremor beyond oral treatment based on EBM‐criteria. Due to the high‐quality evidence for DBS on tremor suppression and improvement in quality of life, the treatment is a good choice for drug‐resistant as well as drug‐responsive but orally insufficiently treated PD tremor. ⁷ , ³³ , ⁵⁷

VIM‐MRgFUS may be an alternative for tremor‐dominant older PD patients with asymmetric tremor presentation as it achieves a relevant unilateral tremor reduction confirmed by lower‐quality evidence. ²⁵ , ⁵⁹ Evidence for long‐term efficacy is increasing, but still limited for MRgFUS, ⁶⁸ , ⁶⁹ whereas for DBS long‐term efficacy is well established. ³¹ , ³²

Potential Limitations

In particular, the biological age limit, multimorbidity resulting in higher surgery risks, and cognitive impairment may restrict DBS accessibility for older PD patients. ⁷ , ⁴⁷

On the contrary, VIM‐MRgFUS may be an option for multimorbid patients who cannot receive DBS. Patients suffering from preexisting gait impairment, for example, caused by peripheral neuropathy, orthopedic causes, or gait problems due to PD, may experience worsening gait problems after DBS. ⁷⁰

Clinical Reasoning

If elderly patients are no longer eligible for DBS, VIM‐MRgFUS can be an alternative, particularly for patients suffering from a predominantly unilateral tremor.

Patients with Impulse Control Disorders

A common adverse event of dopaminergic medication is ICDs. ⁷¹ The prevalence of the 4 main ICDs (hypersexuality, pathological gambling, binge eating, and compulsive shopping) in the general PD population is about 14%. ⁷² ICDs are often linked to the intake of dopamine agonists with D3‐receptor affinity but can also be facilitated by pulsatile l‐dopa treatment alone or the combination of both. ⁷³ , ⁷⁴

Clinical Profile and Therapeutic Aim

ICDs usually occur in the course of a dopamine replacement therapy, especially with dopamine agonists. ⁷³ Risk factors include male sex, younger age at disease onset, and longer duration of the intake of dopaminergic treatment. ⁷³ The main goal for this advanced PD cohort is the reduction/discontinuation of dopamine agonists and—if possible—reduction in l‐dopa to reduce ICDs as their side effects. ⁷⁴

GL, Lower‐Quality Evidence, and Long‐Term Observations

One RCT reported an improvement in hyperdopaminergic behavioral disorders, in general, after STN‐DBS compared to BMT ⁷⁵ ; however, the reduction in ICD subscores was not significant because the study was underpowered for this outcome. ⁷⁴ Lower‐evidence trials report an improvement in ICDs after STN‐DBS in the short‐ and long‐term periods of up to 10 years ⁷⁴ , ⁷⁶ ; 95% of patients with preoperative ICDs showed a symptomatic improvement mainly explained by the reduction in dopamine agonist dose. ⁷⁴ , ⁷⁷ De novo onset of ICDs after STN‐DBS is quite rare, however, mostly due to electrical stimulation of adjacent structures and postoperative adjustments of medical treatment that emphasizes the necessity of perioperative management. ⁷⁷ , ⁷⁸ , ⁷⁹ In contrast, GPi‐DBS is not reported to improve ICDs probably due to the lack of a substantial decrease in dopaminergic medication. ⁷⁴

Evidence level on pump therapies for ICDs is equally low. However, 2 uncontrolled prospective studies on LCIG report an improvement in ICDs after 6 months and 3 years of LCIG treatment in 62 and 8 patients. ⁷¹ , ⁸⁰ The effects are explained by the reduction in pulsatile therapy in favor of a continuous administration. ⁷⁸ Also, CAI therapy is reported to improve ICDs, however, leading to low rates of new‐onset ICDs. ⁷⁸ , ⁷⁹ , ⁸¹ Apart from the reduction in pulsatility by continuous administration, the reduction may also be explained by a distinct dopamine receptor affinity profile of apomorphine. ⁴² , ⁷⁸

Potential Limitations

Inclusion criteria and adverse events for the different therapeutic options are outlined in part 1 (see above: the severely fluctuating patient). Severe psychiatric disorders are a contraindication for all invasive treatments, and often also patients with mild neuropsychiatric symptoms were excluded in trials. ⁸² Therefore, the evidence level for invasive treatment options for patients with ICDs is very low.

Clinical Reasoning

Based on promising results from lower‐quality trials on improving ICDs by reducing dopaminergic medication or its pulsatility, DBS and pump therapies can be considered for PD patients with ICDs on a case‐by‐case basis. However, an elaborate preoperative assessment for severe neuropsychiatric symptoms must be performed to reveal neuropsychiatric manifestations as an absolute contraindication. Moreover, the necessity for a complex perioperative drug management and physicians should sensitize patients to comply with drug withdrawal. ⁷⁵ , ⁷⁸

Patients with Hallucinations and Borderline Psychosis

Psychotic symptoms are nonmotor symptoms frequently observed in advanced PD (~25%–30%). ⁸³ The most common psychotic symptoms are delusions and visual hallucinations (VH) and occur mostly several years after the initial motor onset depending also on patient age. ⁸³ , ⁸⁴ Psychotic symptoms are associated with nursing home placement and an increased carepartner burden. ⁸³ The underlying pathophysiology of VH in PD is still unclear. ⁸⁵ Recent research has improved our understanding of the development of VH, emphasizing an association with cognitive decline and epileptiform discharges in the context of a more diffuse neurodegenerative process. ⁸⁵ , ⁸⁶ Nevertheless, dopaminergic medication is still the most important provoking factor for VH. ⁸⁴ First‐line treatment is reduction in dopaminergic medication and the usage of the antipsychotics clozapine or quetiapine. ⁸⁴ , ⁸⁵

Clinical Profile and Therapeutic Aim

PD patients with VH and borderline psychosis generally suffer from a high disease burden due to their advanced state of disease. Because of the link between dopaminergic medication and VH, the therapeutic aim for patients with VH is the reduction in the dopaminergic dosage and the improvement in other PD symptoms without worsening psychotic symptoms.

GL, Lower‐Quality Evidence, and Long‐Term observations

Although acute severe psychosis is a contraindication for most invasive treatment options, ²⁴ , ⁴⁵ , ⁴⁷ , ⁸² it is theoretically reasonable to use dopamine‐reducing interventions like DBS for improving the VHs without compromising mobility. As such, a large prospective observation trial of 173 patients documented an improvement in hallucinations after 6 months of treatment with STN‐DBS, CAI, or LCIG. ⁸⁷ Further, smaller observational studies report an improvement in VH or at least no deterioration using STN‐DBS, LCIG, or CAI therapy. ⁷¹ , ⁸⁸ , ⁸⁹ , ⁹⁰ , ⁹¹ , ⁹² , ⁹³ , ⁹⁴ Interestingly, CAI is reported to reduce VH despite belonging to the group of dopamine agonists. This is hypothesized to be due to a differing chemical structure and receptor affinity compared to other dopamine agonists. ⁸² , ⁹² However, there are also rare cases of psychotic adverse events for DBS, LCIG, and CAI but also in the BMT groups. ¹⁵ , ²² , ²⁴ , ³⁴ VHs are further reported to be a risk factor for cognitive decline after DBS. ⁹⁵ These findings highlight the difficulty in differentiating between disease‐ and invention‐induced psychotic symptoms.

Overall, only limited recommendations can be made ⁹⁶ , ⁹⁷ ; particularly, clinically obvious l‐dopa‐depending psychosis may be considered for treatment with l‐dopa ‐sparing interventions.

Potential Limitations

Potential limitations are the lack of evidence and potential adverse events. Adverse events are outlined in part 1. As VH predicts a cognitive decline in PD patients even after DBS treatment, pretesting the cognitive status is obligatory for any invasive treatment option. ⁸⁵ , ⁹⁵ A reduced cognitive status is associated with a limited ability for device management in pump therapies. ⁹⁸ , ⁹⁹ Therefore, the prospective disability must be considered in patients with VH, and invasive treatment options might be restricted in the case of nonexistence of care partners. ⁸² , ⁹⁸

Clinical Reasoning

Due to the lack of evidence, decision‐making for the individual patient should be guided by the symptom profile, which can be improved by the different treatment options. Nevertheless, the scarce literature reveals at least no negative effect of invasive treatments on psychosis and VH. Therefore, mild psychotic symptoms should not limit the accessibility to invasive treatment options if otherwise indicated, and dopamine‐sparing interventions can be considered for clearly dopaminergic‐dependent VH and psychosis.

Patients with Minimal Cognitive Impairment or Mild Dementia

PD dementia (PDD) is a common symptom in the late phase of PD; for example, 83% of PD patients were diagnosed with dementia after 20 years. ¹⁰⁰ Typically, cognitive impairment develops over years along a continuum from slight frontal cognitive disturbances to mild cognitive impairment (MCI) and to mild and advanced PDD. ¹⁰¹

Clinical Profile and Therapeutic Aim

PD patients seeking invasive treatment commonly present with burdensome motor symptoms. Therefore, the therapeutic goal is the improvement in motor symptoms, and there is no evidence that cognitive symptoms can be improved using invasive therapy. ¹⁰² , ¹⁰³ , ¹⁰⁴

GL, Lower‐Quality Evidence, and Long‐Term Observations

There are no controlled or uncontrolled trials with MCI or dementia patients. Long‐term cohort studies in DBS therapy suggest sustained efficacy on motor symptoms without any effect on cognition. ³² , ¹⁰² The occurrence of PD‐related late‐stage disability milestones is unlikely to be worsened but rather slightly improved with DBS. ¹⁰⁵ One study in 79 PD patients with a mean age at surgery of 59 years found half of the patients developing dementia and 42% living in nursing homes on follow‐up after 10 years. The only predictor for nursing home placement was older age at surgery, implying an overall loss of cognitive and physical capacity. ¹⁰⁶ Another study involving more than 100 patients with STN‐DBS found a prevalence of dementia of 29.8% after a 10‐year stimulation. ¹⁰⁷ Predictive factors for dementia were age and frontal score at the time of surgery and male sex. MCI on the contrary does not alter the prognosis after DBS surgery compared to PD patients without cognitive decline although preoperative MCI constitutes certainly a risk factor for developing dementia later. ¹⁰³ , ¹⁰⁸ For LCIG and CAI, no data are available on patients with incident dementia after treatment was initiated. Not surprisingly, mortality in patients with LCIG therapy depended on the cognitive status at treatment onset in a 10‐year study. ⁴⁵ , ¹⁰⁹ , ¹¹⁰

Potential Limitations

Many experts consider dementia as a contraindication for DBS because the patient's decision‐making may be restricted, the life risk associated with DBS implants may be increased, and the therapeutic benefit may be lower for PDD due to therapy‐resistant end‐stage symptoms, such as axial symptoms, dysphagia, or dysarthria. However, most patients developing dementia while stimulated want their battery replacement or would not tolerate switching the stimulator off. MRgFUS can be considered for patients with severe medication‐resistant tremor, although not yet unconditionally recommended. ⁷ Side effects, but also intra‐procedural cooperation, may be limiting factors. For LCIG and CAI pump therapy a reduced cognitive status may limit the ability for device management. ⁹⁸ , ⁹⁹ Indeed, technical problems, especially the potentially dangerous accidental removal of the inner tube, are more frequently observed in demented patients. ¹¹⁰ , ¹¹¹ , ¹¹² Therefore, pump therapy may be considered in mild and moderate dementia when care partners supervision is provided, but it can increase the care partners' burden in the advanced state of disease. ⁸² , ⁹⁸

Clinical Reasoning

Both controlled and cohort studies are tremendously lacking. Merely case reports and observational studies report an improvement in motor symptoms in PDD patients due to interventional therapies even in the long term. ⁹⁰ , ¹¹⁰ Therefore, well‐selected patients with mild dementia or MCI may represent borderline candidates for DBS treatment if they suffer from severe fluctuations. ¹¹³ An alternative for patients with an asymmetric therapy‐resistant tremor constitutes MRgFUS keeping the limitations in mind. Pump therapies are the other option sometimes used for this patient group, but its use critically depends on patients' cooperation. Main factors for decision are the severity of motor and neurobehavioral symptoms as well as the caregiving environment. For new implantations for such patients, the trade‐off between unbearable motor symptoms and handicap due to dementia should be very clear in favor of motor disease.

Patients in Palliative Care

The palliative care aim of the World Health Organization is improvement in quality of life of patients, who suffer from a life‐threatening disease, as well as of their families and carepartners. ¹¹⁴ It was suggested that PD disease progression can be divided into 4 phases, ceasing with the palliative care phase in the last 2.2 (±2.2) years. ¹¹⁵ Palliative care in PD has been implemented in clinical practice and scientific research just in recent years. ¹¹⁶

Clinical Profile and Therapeutic Aim

Palliative PD patients suffer from burdensome motor and nonmotor symptoms, in particular slowness, stiffness, rigidity, pain, anxiety, fatigue, memory issues, speech problems, or dysphagia. ¹¹⁵ , ¹¹⁷ However, some patients received an interventional therapy before reaching the final phase of their disease. But later‐stage PD is often related to l‐dopa‐resistant motor symptoms, for example, axial symptoms, which also interventional therapies cannot improve sufficiently leading to the question when to terminate the interventional therapy. ¹¹⁸ , ¹¹⁹ , ¹²⁰ Therefore, before terminating the device‐aided therapy, a careful assessment is mandatory to confirm the benefit which the patient still has. ¹¹⁹ Overall, the superordinate goal for palliative patients should always be the improvement in quality of life. ¹¹⁶

GL, Lower‐Quality Evidence, and Long‐Term Observations

Higher‐quality evidence in palliative PD patients treated with device‐aided therapies is lacking. However, a small cohort study compared the quality of life of palliative care PD patients with DBS or LCIG therapy to BTM and could not show any significant difference. ¹²¹ Nevertheless, if the existing treatment improves the individual's quality of life, device‐aided therapies should be continued. ¹¹⁹ The treatment can be seen as a comfort care when motor treatment–responsive PD symptoms are the reason for a diminished quality of life. ¹¹⁹ As DBS is not reported to prolong the dying process, the treatment does not contradict the palliative care concept. ¹¹⁹

Potential Limitations

As many patients suffer from communicative and cognitive impairments, decision‐making is difficult for palliative PD patients. Therefore, the decision has to involve family and care partners who also experience a disease burden and whose lives are equally influenced by the interventional therapy. ¹¹⁵ , ¹²² , ¹²³ , ¹²⁴ To this end specific educational toolkits have been implemented following international GLs. ¹²⁵

Clinical Reasoning

Due to the lack of studies, conclusions are not possible. The termination of the treatment should be dependent on the individual response of the patient to the treatment and the value of the individual benefit for the patient's life quality. It has been proposed to consent on treatment withdrawal already at the start of the interventional therapy to finally “fulfill the patient's wish,” ¹¹⁵ , ¹²⁴ but experienced clinicians have doubts if such complex decisions can be anticipated.

Conclusion

Invasive treatment options are an alternative for PD patients suffering from a disease burden inadequately managed by oral treatment. Due to the new EAN/MDS‐ES GL, on which this paper focuses, the evidence for each therapy is well documented. Therefore, the current paper outlines factors, which have to be considered within the decision‐making process, for 8 typical PD symptom profiles. However, physicians and patients should discuss each option carefully focusing on the individual profile and considering the pros and cons of any therapy alternative.

Author Roles

(1) Project: A. Conception, B. Organization; (2) Figure: A. Design, B. Execution, C. Review and critique; (3) Manuscript preparation: A. Writing of the first draft, B. Review and critique.

D.B.: 1A, 1B, 1C, 2A, 2B, 3A

K.S.: 1A, 3A, 2C, 3B

S.P.: 1A, 2C, 3B

E.M.: 2C, 3B

A.A.: 2C, 3B

G.D.: 1A, 1B, 2A, 2C, 3B

Disclosures

Ethical Compliance Statement: This review was conducted in accordance with all ethical principles. Informed patient consent was not necessary for this work. We confirm that we have read the journal's position on issues involved in ethical publication and affirm that this work is consistent with those guidelines.

Funding Sources and Conflicts of Interest: None of the authors have a conflict of interest to report regarding this paper.

Financial Disclosures for the Previous 12 Months: Dana Brinker and Katarzyna Smilowska repot no disclosures. Steffen Paschen received study grants, lecture fees, and travel expense reimbursement from Insightec GmbH, Boston Scientific GmbH, and Medtronic GmbH. Angelo Antonini has received compensation for consultancy and speaker‐related activities from UCB, ConvaTec, Infusion Care, Bayer, General Electric, Britannia, AbbVie, Zambon, Bial, Theravance Biopharma, TreeFrog Therapeutics, Roche, and Medscape. He receives research support from Bial, Lundbeck, Roche, Angelini Pharmaceuticals, Horizon 2020 Grant 825785, Horizon 2020 Grant 101016902, Ministry of Education University and Research (MIUR) Grant ARS01_01081, Cariparo Foundation, and Movement Disorders Society for NMS Scale Validation. Elena Moro reports consultancies for Abbott, Medtronic, and Kyowa and major grants from Medtronic and Ipsen. She received payment for lectures from Abbott and Medtronic. Günther Deuschl has served as a consultant for Boston Scientific, Cavion, and Functional Neuromodulation. He has received royalties from Thieme Publishers and funding by the German Research Council (SFB 1261, T1).