Parkinson’s disease DBS: what, when, who and why? The time has come to tailor DBS targets

Brain targets (ventral intermediate thalamus, STM, GPi, PPN, centromedian thalamic nucleus, zona incerta & others)

A plethora of new information has been emerging that will assist clinicians in selecting the most appropriate brain target for a given PD patient. Early in surgical therapy, many DBS groups implanted the ventral intermediate (VIM) thalamus. However this approach gradually fell out of favor with the realization that VIM was most effective for upper extremity tremor, with much less measured efficacy for the other cardinal motor features of PD (e.g., bradykinesia and rigidity). However, very select patients with upper extremity tremor and without other prominent disabling features of PD are still considered to be appropriate for VIM DBS [13].

Following an era of encouraging results in ablative/lesional brain therapy, first in the VIM thalamus (thalamotomy) and then in the GPi (pallidotomy), the GPi emerged in the early 1990s as a potential target for DBS therapy. The DBS approach was preferred over lesion therapy by most centers, especially those interested in bilateral therapies. The reason for the preference was that bilateral lesions had been commonly associated with unacceptable speech, swallowing and cognitive deficits. However, the GPi target was quickly overtaken by STN DBS [9,14], following the revelation that STN lesions were found to be very effective for the parkinsonian primate [15] and that drastic medication reduction could be realized in some human cases [16]. However, recent trials have demonstrated that GPi DBS is also an effective target for treatment of the levodopa-responsive symptoms and the motor fluctuations associated with PD [4,7]. Small individual differences in motor, mood, cognitive, adverse event and quality of life outcomes have been observed across available comparative studies (e.g., bradykinesia, depressive symptoms, apathy, verbal fluency, falling and speech issues), and more secondary analyses of the respective databases are needed.

There are potential ways to interpret emerging DBS outcome data to maximize benefits and to minimize risks. An interdisciplinary DBS team (neurosurgery, neurology, neuropsychology and psychiatry) can weigh the benefits and risks of choosing an individual target for a potential DBS patient. Figure 1 shows the pathway at the University of Florida (FL, USA) for the interdisciplinary ‘fast track’ work-up, as well as the details of the follow-up interdisciplinary discussion for any potential DBS patient. We refer to ‘fast track’ at the University of Florida as an interdisciplinary process where patients can be seen and worked-up by multiple specialties on consecutive days. The work-up is then followed by an interdisciplinary discussion. We understand that not all centers have psychiatrists on their DBS teams, and in cases where a psychiatrist is not available, screening is usually performed by a neuropsychologist. In our experience, we have found the psychiatrist and the psychiatry evaluation to be essential to a complete work-up.

There has been a recent surge in interest in pedunculopontine nucleus (PPN) DBS to address levodopa-resistant PD-related gait and balance dysfunction [17,18]. However, there is a paucity of data available to recommend PPN DBS, and thus it is currently limited to the research setting. Important issues that remain to be resolved for PPN DBS are the inclusion/exclusion criteria (on vs off freezing, gait problems, balance problems and combinations of clinical issues), locating the best target region (imaging and physiology), whether preoperative white matter lesions on MRI impact outcome (comorbidities), and to better define and clarify which motor and non-motor features of PD may respond to this approach. Furthermore, it is important to note that Stefani et al. may have inadvertently targeted the peripeduncular nucleus and not the PPN [18,19]. Other recent experimental brain targets for PD DBS have included the centro-median thalamic nucleus [20] and the zona incerta [21], but similar to PPN DBS, these targets remain investigational with little data available to guide clinician–patient discussions.

Unilateral versus bilateral operations

There are no large-scale studies available that directly compare a unilateral versus bilateral approach to PD DBS. The randomized COMPARE DBS cohort data revealed that unilateral STN and unilateral GPi DBS targets were both efficacious in addressing motor dysfunction, however, they had similar but slightly different mood and cognitive side-effect profiles [7]. There remains an open question as to whether a unilateral approach would be a safer and more desirable option for a subset of patients.

An interesting follow-up study from the COMPARE cohort suggested that unilateral GPi DBS had better quality of life outcomes when compared with unilateral STN DBS [12], raising the question that if patients are to be implanted unilaterally, GPi may prove to be a more appropriate target. A follow-up study currently in press revealed that 21 out of the 44 patients (47%) from the COMPARE cohort remained unilateral, and were followed for an average of 3.5 years after surgery [22]. Of these cases, 14 out of 21 (67%) had a GPi target. The most common reason for adding a second-sided DBS was inadequacy in addressing motor symptoms. Satisfaction with motor outcomes was the most common reason for remaining unilateral. Those who chose a second DBS lead had significantly higher baseline UPDRS III motor scores and a significantly lower asymmetric index (meaning that their symptoms were rated as closer to equal on both sides of the body). The logistic regression analysis revealed the odds of proceeding to bilateral DBS was 5.2-times higher for STN compared with GPi. For every 1% increase in asymmetry, the odds of bilateral DBS increased by 0.96 [22]. It is unknown why unilateral GPi DBS may be better in select cases than unilateral STN DBS, but part of the explanation may lie in the mechanisms of dyskinesia suppression (medication reduction in STN vs a more direct suppression of dyskinesia in GPi). The differences between targets will need more scrutiny, as it is likely that STN also has some dyskinesia-suppressing properties.

Simultaneous implantation into two brain sides versus a staged approach

When considering a simultaneous versus staged approach to DBS, some explanation of the different approaches is required. Two DBS leads may be placed in the same intraoperative sitting (simultaneous implantation), or alternatively they may be separated by days, weeks or months. Similarly, the batteries (impulse generators) may be placed on the same day, or alternatively placed days, weeks or even a month following lead insertion. Although there is no data to support the utility for timelines of these various approaches, some DBS interdisciplinary teams feel that extended intraoperative time may be an important factor in increasing complications, particularly in the elderly patient (>70 years old), the frail patient or in the patient with multiple comorbidities. Some teams have also begun to exercise caution in operating on patients over 70 years of age, although this point is highly debatable among the experts.

The ‘pros’ supporting a single intraoperative bilateral lead insertion session include patient convenience, eliminating a second and same-day operation and/or an economic advantage for the healthcare system. The ‘cons’ include longer intraoperative times, the potential for greater complications (including intra- and post-operative mental status changes), and in some countries, decreased compensation for performing DBS in a single intraoperative sitting. It has been hypothesized that a unilateral electrode implantation may be less likely to result in mental adverse effects compared with bilateral electrode implantation due to microlesioning, but further studies are needed to confirm such suspicions. When it comes to battery placement, many surgeons prefer to delay placement for a few weeks and even up to a month in order to limit postoperative confusion, which may occur by prolonging the lead insertion surgery. A delay has the added benefit of allowing brain edema to resolve prior to attempting device programing. It should also be noted that in centers not allowing MRI following a DBS lead insertion, this may impact the ability to place a second contralateral DBS. There is an ongoing debate as to the safety of MRI post-DBS, with many groups, including the National Parkinson Foundation, citing an excellent safety record with MRI and DBS, especially when performed under optimal conditions [23].

Recently, a study was published that explored postoperative stays following unilateral DBS surgery. This study examined factors potentially related to extended stays. Most patients in the 115 patient cohort (79%) had a hospital stay of 1 day following surgery. The most frequent reasons for delayed discharge (>1 day) included mental status change (n = 6) and hemorrhage (n = 5). Those with delayed discharge had significantly lower pre-surgical cognitive screening scores (Mini-Mental State Evaluation [MMSE]), higher presurgical ‘on’ medication motor score and more microelectrode passes compared with the immediate discharge group. In correlation analyses, increased length of hospital stay was significantly associated with more microelectrode passes, higher presurgical ‘on’ medication motor scores and decreasing MMSE scores. When the significant variables from the preliminary analyses were entered into a Poisson regression model, a greater number of microelectrode passes, as well as lower MMSE scores, remained significant predictors of increased length of stay. These issues should be more closely examined in larger studies, studies including a bilateral simultaneous approach and studies also including long-term outcomes [24].

Closed-loop technologies & Parkinson’s disease (improving management)

There are currently no US FDA-approved DBS devices that allow for either scheduled stimulation or responsive (closed-loop) stimulation. There are limited available settings on devices that allow for cycling of current, and these settings have been largely unexplored. Recently, a device has been introduced for epilepsy patients, and there is hope that similar devices may play a role in the management of PD [25]. One reason for the enthusiasm for such devices is that in the PD brain, regions have been identified that predictably oscillate at frequencies such as the β-band. New devices may have the potential to detect this β-band frequency (and other frequencies during rest or during action). These frequency oscillations may prove useful in choosing programing parameters for chronic DBS management [26]. Though closed-loop technologies are interesting, it should be noted that there may be cumulative motor effects from long-term DBS [27].

What is the current rationale for when to operate?

Deep brain stimulation for PD is currently offered to patients with medication refractory on–off fluctuations, dyskinesia or tremor. These patients may be selected as candidates following an extensive interdisciplinary screening. The screening should aim to elucidate which features of PD are most responsive to dopaminergic therapy (responsive features tend to have the best response to DBS, with the exceptions of tremor, dyskinesia and motor fluctuations), it should identify cognitive/psychiatric issues, and it should assess comorbidities that may increase the risk or limit the effectiveness of therapy [8]. The major unanswered question for DBS therapy has been when to say enough is enough, and to proceed to DBS therapy [28]. Most practitioners agree that when medication intervals become very close in time (within 2–3 h), and on–off fluctuations, dyskinesia or tremor emerge and are difficult to control, then it is time to at least consider the use of DBS therapy. Several recent studies have revealed an advantage of DBS over medical therapy, including the UK PD surge trial [6], (which allowed apomorphine therapy in medication and in DBS groups), the VA Cooperative Study [4] and the NEJM Quality of Life Study [3].

Should the timeline for operating be moved up?

Traditionally many expert PD centers have delayed invasive surgical therapy for as long as possible. This purposeful delay has raised the question, ‘how long is too long?’ DBS has not been shown to have a disease-modifying effect. As DBS is not disease modifying, to proceed to surgical therapy symptoms should, at a minimum, be difficult to manage with medications, and they should have a profound impact on the sufferer’s quality of life. If patients are still working, and DBS has the potential to keep them working, this should be considered in the decision-making process. To date, studies have not shown any advantage for DBS to keep PD patients at work. These types of studies are, however, very complex to perform, and very complex to interpret [29,30], as there are many factors as to why a patient may not return to work.

The many ways to define early DBS & how early is too early?

A recent study suggested that DBS in less advanced cases (lower UPDRS scores) may be feasible and efficacious [29,30]. Trials published in the literature to date have not included patients with short disease durations (1–4 years) or patients without motor fluctuations. The neuroprotective (or disease-modifying) effects of surgery have not been substantiated, and thus cannot be ethically used to argue for earlier intervention in individual patients. There is a solid argument evolving for cost savings with early DBS. However, arguments based on motor, non-motor and quality of life features have less of an evidence base at this time. Better methodologically constructed and adequately powered clinical trials with carefully conceived end points will need to be performed in order to settle the questions surrounding early DBS intervention.

‘Early DBS’ covers several scenarios. Consider the example of a 60-year-old patient diagnosed with PD 2–4 years prior, experiencing life-altering disability from tremor, despite maximal combinations of levodopa, dopamine agonists and anticholinergics. Most clinicians and multidisciplinary surgical teams would agree that this could be a reasonable case for early intervention.

A more controversial case for ‘early’ intervention would be that of a 60-year-old patient with 2-h levodopa dose intervals, experiencing mild dyskinesia, as well as on–off fluctuations (regardless of disease duration), with minimal off time. The patient might desire DBS, believing the surgery to be inevitable, with the goal of minimizing the surgical risks of hemorrhage and cognitive deterioration. Although not all experts would agree that surgery should be performed in such a case, the scenario does represent a rational argument for earlier intervention.

In our opinion, the most controversial scenarios for ‘early’ intervention are:

• Patients who have not yet received levodopa or standard of care medication therapy;

• Patients who have been diagnosed for less than 5 years;

• Patients without motor fluctuations, or patients just beginning to manifest motor fluctuations;

• Patients seeking to reduce or discontinue medication with short disease durations (<5 years);

• Patients with short disease durations (<5 years) and nondisabling and nonbothersome dyskinesia;

• Patients who desire to continue working and desire DBS in the hope of continuing to do so for a longer time period of time (some centers view this as noncontroversial despite a lack of evidence-based studies);

• Patients with short disease durations (<5 years) seeking improvement in non-motor manifestations [31].

We believe that patients who fall into these above categories should, in most cases, have DBS performed only in the context of an Institutional Review Board-approved clinical research trial.

There are several reasons for not performing DBS routinely in the above situations. There is a higher chance of confusing a Parkinsonian disorder (such as multiple system atrophy) within the first 5 years of diagnosis, and DBS has not conclusively been shown to benefit patients with ‘Parkinson plus’ syndromes (it should be noted that 5 years is an arbitrary time interval and has not been shown to be an adequate cut-off by formal studies). Furthermore, PD is not one disease, but rather a syndrome. Patients with fewer than 5 years of PD may have an insufficient symptom history for proper surgical stratification. For example, some patients with tremor-predominant PD will continue to function well for decades on dopaminergic therapy alone. Currently, we have no way to easily identify such patients who may never need surgery, and therefore we cannot remove them from the surgical candidate pool.

There are ongoing studies of early intervention underway in the USA and in Europe [32]. Although we expect that this data will shed light on the more controversial issues regarding this debate, it is unlikely to settle it. The data from each study will be specific and based on individual study design. Levodopa and the motor UPDRS are currently used as the best preoperative indicator of symptomatic benefit for DBS in PD. They may be less useful in indicators in ‘early’ scenarios.

Designing early intervention trials to create or revise outcome measurements will be challenging. Early intervention trials will be costly, and the follow-up required longer than what we have come to think of as standard for the field. The data from these trials will also be necessary for the formulation of rational criteria under which early intervention can be considered [31].

Many centers have been shifting toward earlier implantation of DBS devices. Physicians should keep in mind that earlier DBS translates to increased potential for adverse events over the lifetime of a patient and therefore to more programing visits and also to more battery changes. Therefore, patient commitment and patient-perceived benefits must also be carefully considered when deciding whether or not to perform surgery at an earlier stage of PD [31]. There are many flaws in existing studies of early DBS, especially in methodology and outcomes, and future studies will need to carefully examine these issues.

Symptom profiles that may or may not benefit from DBS

Data derived from the available randomized DBS studies suggest that PD patients with medication-refractory and difficult to control on–off fluctuations may be the best candidates for DBS therapy [3–7,28]. In the UK PD Surge trial, the authors collected information on the main reasons why patients considered DBS surgery. Severe ‘offs’, dyskinesia, and tremor were the most commonly cited indications [6]. Although it is now well-established that levodopa-responsive symptoms of PD have the greatest response to surgical interventions, the effect of disease progression may not be routinely included in preoperative patient discussions, where risks and benefits can be thoroughly vetted. Discussions with patients should routinely involve the realization that walking, talking and thinking could worsen not only from stimulation- or lesion-induced effects, but also as a result of disease progression [8,33] (these can occur in both unilateral and bilateral lead implantations and it is unknown whether there is an advantage of one approach over the other). Patients with highly asymmetric symptoms may need only a single-sided DBS operation, and in these cases taking a conservative approach to therapy may improve the risk–benefit ratio (i.e., waiting until after a single lead has been placed and medically/surgically optimizing before proceeding to a second-sided DBS) [34–36].

Less is known about the cognitive and mood sequela of DBS surgery. The most common cognitive issue that has emerged has been reduction in verbal fluency (patients may complain they cannot get words out of their mouth). This problem has been demonstrated in multiple studies, including recently by a reliable change index that was employed to be sure the effect was actually DBS specific [37,38]. Interestingly, very few DBS studies have carefully controlled for ‘lesional’ effects and separated them from stimulation-induced effects. In the COMPARE cohort, turning the stimulators on and off at 7 months postimplantation revealed a persistent negative effect (lesional effect) on fluency, particularly in the STN group [7]. Interestingly, stimulation-induced effects were also observed in both STN and GPi, particularly when adjusting the stimulation to more ventral (deeper) DBS contacts [7,39].

The most commonly observed mood effect following DBS surgery has been a slight improvement in depressive symptoms. Recent data has suggested that this improvement may not be persistent, and the VA Cooperative Study Data at 24 months actually suggested slight worsening of mood in the STN DBS group [4]. Interestingly, anger has emerged as a common side effect post-DBS surgery, and has been recently shown to be more of a ‘lesional’ effect of both STN and GPi DBS in PD [7,40].

Several recent papers have cited problems with impulsivity in STN DBS patients and data has revealed that DBS tends to speed up decisions during high-conflict situations. Impulsive and compulsive behaviors, dopamine dysregulation and emergent issues when dual or multitasking have led to a re-examination of the cognitive and behavioral effects of DBS [41–45]. Teasing out these effects may need a more rigorous approach with more complex and more loaded (more difficult tasks for patients) neuropsychological testing.

Finally, one must also consider whether a medication reduction (more commonly seen in STN) may be useful in cases with medication-related psychosis or behavioral issues, which may impact target choice or unilateral versus bilateral implantation.

Interdisciplinary team screening & discussion

The inclusion and exclusion criteria for DBS have continued to evolve for various disorders and syndromes, and they continue to evolve in PD. The first step to ensure success in PD is to select the appropriate candidate. Patient selection for DBS should, in the best case scenario, utilize a multidisciplinary team, and preferably an interdisciplinary team. The preoperative detailed assessment by multiple disciplines should be a prerequisite for surgical consideration. An adequate team usually includes a movement disorders neurologist, a neuropsychologist and a neurosurgeon. Our group also recommends that a psychiatrist be involved if possible, especially since many of the urgent and emergent issues in DBS therapy involve psychiatry. Physical therapists, occupational therapists, speech therapists, social workers and financial counselors should be involved on a case-by-case basis. Each member of the interdisciplinary screening team should independently evaluate the patient and a meeting should be convened to review and discuss findings. The review should be inclusive of the past history, medical imaging studies, and if available, a video examination. Standardized rating scales should be performed (such as the UPDRS, with an on–off dopaminergic evaluation) and comorbidities should be discussed. Candidacy should also be determined based on an individual patient’s desired expectations and goals [8,33,46,47]. Any level of cognitive impairment should be identified and discussed by the interdisciplinary team, as this may affect outcome. The patient should be contacted and apprised of the interdisciplinary discussion(s). Figure 1 summarizes the interdisciplinary University of Florida ‘fast track’ process.

Development of a patient-centered outcome

The focus of research on outcomes from medical procedures has shifted in recent years from purely objective scales to an emphasis on patients’ perceptions of potential outcome(s). The majority of clinicians, patients and caregivers report improvements following DBS surgery. Reports of improvement derived directly from questions asked to patients, clinicians and caregivers often fail to correlate. In addition, patient perceptions of outcome are often associated with non-motor factors, including anxiety, perceived social support and nonphysical domains of quality of life. Patients’ perceptions may thus be less associated with motor and physical factors. Clinicians’ perceptions seem to be more closely tied to motor symptoms and other physical factors. Finally, caregivers’ perceptions may not be correlated to many of the outcome variables we typically measure. The differences between patients [29,48], clinicians and caregivers all underscore the importance of anxiety, perceived social support and other non-motor factors as important considerations in preoperative DBS discussions [49].

When to say enough is enough in fluctuating Parkinson’s disease patients

The increasing use of DBS for PD has drawn the attention of both the medical and lay communities and has resulted in an appropriately increased scrutiny of outcomes. We have celebrated successes but also been challenged by patients who have undergone surgery with results falling short of expectations – a group now referred to as ‘DBS failures’ [10]. We now have a developing evidence base that compares DBS with standard medical therapy. The large VA Cooperative multicenter study [50], along with the NEJM Quality of Life Study [3] and the UK PD Surge Trial [6], have helped us to appreciate the potential benefits of DBS over best medical therapy. The improvements in ‘on time’ of 4–6 h have helped us draw a better picture of when “enough is enough with medicines, and when it is appropriate to proceed to DBS surgery” [28]. The results of all three of these studies collectively suggest that DBS is superior to best medical management for the treatment of on–off fluctuations in well-selected moderate-to-advanced PD patients. We addressed some of these issues in a recent article, concluding, that the “general neurologists and general practitioners should refer moderate to advanced on–off fluctuating PD patients to experienced surgical centers for complete interdisciplinary screening for DBS surgery. Although with the current technology many PD patients may not be appropriate surgical candidates, DBS surgery carried out on the right patient by an experienced team can provide magnificent benefits. These benefits, we now understand, may extend beyond what can be achieved with medicines alone, and this information should give practitioners solace when saying enough is enough” [28].

Exceptional cases: medication-refractory dyskinesia & tremor in otherwise less than optimal candidates

In optimal circumstances, potential PD DBS candidates report to their neurologist’s office 12 h following cessation of their dopaminergic medications. Patients are then evaluated with a UPDRS scale in the ‘off’ state. Afterwards, they are challenged with a suprathreshold dose of dopaminergic medication and reevaluated in their best ‘on’ state. Optimal surgical candidates typically demonstrate at least a 30% improvement in the motor portion of the UPDRS [51]. However, there may be patients who fail to achieve a 30% improvement with levodopa or dopaminergic medications, but have alternative and potentially responsive indications for DBS (dyskinesia, ‘on/off’ motor fluctuation and medication-refractory tremor), some of whom may not fulfill cognitive criteria to receive DBS surgery. Each of these cases should be individually evaluated by the interdisciplinary DBS team to determine the potential for surgical candidacy. Medication-refractory tremor and severe dyskinesia may, in select cases, be appropriate indications for DBS candidates. In cases where cognitive issues may otherwise negate surgical candidacy, many centers have been utilizing unilateral GPi DBS, although more data will be required on this subject in order to assess the efficacy and safety of this approach [8,10].

It should be emphasized that the UPDRS is a clinical research tool that should not be used in isolation. Insurance companies and interdisciplinary teams should not deny DBS based on less than 30% improvement or on UPDRS on or off scores that do not exceed arbitrary limits (such as a common cut-off score of 30). Each patient should be individually evaluated for appropriateness of DBS.

DBS versus apomorphine, duodopa, stem cell trials & gene therapy trials

There is little available data comparing DBS to other surgical interventions, such as apomorphine pumps, duodopa pumps or to stem cell and gene-therapy approaches. The lack of data often leaves the clinician in a difficult position when attempting to compare standard versus investigational therapies for an individual patient. The UK PD Surge trial did provide some limited data suggesting that in patients still having difficulties post-apomorphine pump placement, DBS may represent a reasonable approach [6]. Future comparative trials, especially of apomorphine and duodopa therapy (compared with DBS) may help to elucidate which phenotype of patients may be the most appropriate for each type of therapy. The lack of data and lack of approval beyond research for stem cells and gene therapy make comparison with DBS difficult.

Key issues in maximizing DBS benefit

The maximizing of DBS benefits may be influenced by target selection and by laterality.

Key issues in minimizing risks

The minimization of risks may also be closely associated with target choice and may be associated with whether unilateral or bilateral DBS surgery is chosen for an individual patient.