**1. Introduction**

Despite having being introduced in clinical practice more than 20 years ago, selection criteria for deep brain stimulation (DBS) as an effective treatment for advanced Parkinson's disease (PD) still rely on the 'Core Assessment Program for Surgical Interventional Therapies in Parkinson's Disease' (CAPSIT-PD) published in 1999 [1]. These criteria were primarily designed to facilitate clinical research, harmonizing the cohorts of clinical trials. However, most of the indications provided in the CAPSIT-PD document were introduced as guidance into the clinical practice of DBS centers worldwide, being extremely useful in supporting the selection of candidates [2]. Twenty years later, these indications could be considered both restrictive and out-of-date, because the knowledge on PD progression, phenotyping, and genotyping has strongly evolved over the last 20 years. Indeed, according to CAPSIT-PD, only 1.6% of PD subjects would be eligible for DBS, rising to 4.5% when applying more flexible criteria [3].

Moreover, a growing number of studies have reported novel data on the outcome of DBS in the short- and long-term follow-up and proposed predictors of DBS response [4,5]. However, evidence on how to improve and refine the selection process based on these insights for candidates to DBS is still

lacking. Finally, despite the consolidated e fficacy of DBS in improving PD cardinal symptoms and motor complications, the factors predicting a successful outcome on activities of daily living (ADL) and quality of life (QoL) have been addressed only by a few studies so far [6].

Advances in understanding the heterogeneity of PD presentation, courses, phenotypes, and genotypes impose a better identification of DBS candidates as a research priority. Additionally, DBS invasiveness, cost, and the possibility of serious adverse events make it mandatory to predict as accurately as possible the clinical outcome when informing the patients about their suitability for surgery.

Here, we appraised the DBS pre-surgical assessment for PD starting from the original CAPSIT-PD document and addressed the following topics which may impact on the selection process: early versus delayed DBS; the evolution of the levodopa challenge test; the relevance of axial symptoms; new focus on patient-centered outcome measures; the relevance of non-motor symptoms; and a new role for genetics. Our main aim was to highlight current pitfalls and potentialities in the DBS selection process, stimulating future randomized control trials (RCT) to address specific needs.

### **2. Early Versus Delayed DBS: How Early?**

### *2.1. The Standard Rule*

The CAPSIT-PD document recommended that a patient considered for interventional surgery should have a diagnosis of idiopathic PD and a minimum disease duration of five years [1]. These requirements were developed to exclude people with atypical parkinsonism, given the absence of benefits and the risk to harm patients with no idiopathic PD [7].

### *2.2. Pros and Cons*

The concept of a five-year disease duration has been challenged upon the results of a large RCT published in 2013 (the EARLY-STIM trial) [8]. In this trial demonstrating the superiority of subthalamic (STN) DBS compared to medical therapy alone, patients were included when having a PD diagnosis of ≥4 years, and fluctuations or dyskinesia present for four years or less [8].

The EARLY-STIM trial endorsed a conceptual change about the use of DBS for PD favoring a paradigm shift from DBS as the last therapeutic option for advanced disease stages toward an earlier approach for patients experiencing motor complications. This paradigm change is based on three relevant points: (1) the confirmation of DBS safety over the years, even in the long term; (2) the grea<sup>t</sup> efficacy of DBS in improving the QoL of patients, even superior to levodopa alone [9]; (3) an earlier intervention could preserve functional capacity. The evidence of e fficacy provided by the EARLY-STIM trial led the U.S. Food and Drug Administration (FDA) to extend the DBS indication to patients with a four year PD diagnosis in the presence of at least four months of uncontrolled motor complications [10].

The EARLY-STIM trial has triggered discussion as to whether its findings should be translated into clinical practice [11]. Firstly, the shorter disease duration at the time of surgery might pose the risks of including subjects with atypical parkinsonism for which the five-year rule has been developed for CAPSIT-PD. However, in the EARLY-STIM cohort, the mean disease duration of the surgically-treated group was 7.3 ± 3.1 years and only three cases (0.8% of the cohort) were re-diagnosed as non-idiopathic PD eight years after the first randomization and approximately 15 years after diagnosis [12]. However, it should be noted that the issue of a shorter disease duration at the time of surgery might mirror a greater and faster burden of disability which is also associated with specific genotypes associated to PD, such as severe and complex glucocerobrosidase (*GBA*) gene variants [13], which have been associated to poor DBS functional outcome [14].

A related matter is the di fficulty in predicting the trajectory of disease progression at such an early stage, either for more benign or rapidly progressing phenotypes [15], with the consequent risk of referring to surgery patients whose motor fluctuations could remain mild for a long time, or vice-versa, those who may develop symptoms non-responsive to DBS and severely impacting ADL and QoL. However, when looking carefully into the EARLY-STIM cohort, all patients had experienced either motor or psychiatric disability due to PD. Accordingly, further analyses on this population have demonstrated that STN-DBS was successful in improving freezing of gait (FoG) in OFF medication condition [16], which a ffected 52% of the patients at baseline. Remarkably, behavioral complications linked to dopaminergic overmedication had a better outcome in the neurostimulation group [17].

How early should DBS be considered in PD? A pilot open label study on 28 patients suggested to consider DBS even earlier, before motor complications arose [18]. However, despite long term follow-up data on the same cohort [19], the impact of early surgical intervention in such earlier stages is still unknown and should be carefully interpreted including the risks we mentioned above but also a presumptive neuroprotective e ffect [20]. Finally, it should be taken into account that DBS (STN-DBS in particular) allows a reduction in dose of dopaminergic therapies in most patients [21]. Although favoring the improvement of dyskinesia, the reduction in antiparkinsonian drugs could have role in improving impulsive-compulsive behaviors and obsessive-compulsive and paranoid traits [22,23].
