Parkinson's Disease in 2026: When and How to Choose Advanced Therapy

Bottom Line: Treatment of Parkinson’s disease is best approached as a staged problem. In early disease, the questions are when symptoms justify treatment and which class best fits the patient; levodopa should be started when symptoms warrant it, since the delayed-start LEAP trial (NEJM 2019) found no difference between early and delayed initiation, arguing against both neurotoxicity and disease modification. In the middle years, management becomes a problem of pharmacokinetics — reducing OFF time and limiting dyskinesia. In advanced disease, the central decision is timely referral for device-aided therapy across three broad modalities — deep brain stimulation, MR-guided focused ultrasound, and continuous infusion — before frailty, cognitive decline, or psychiatric instability narrow the options. Levodopa remains the most effective symptomatic therapy, and no available treatment has been shown to slow disease progression.

The Treatment Arc in Parkinson’s Disease

Parkinson’s disease evolves through recognizable therapeutic phases. Early in the course, most dopaminergic agents provide meaningful symptomatic benefit, and the practical question is how much symptom control is needed now rather than how long levodopa can be deferred. Later, as presynaptic buffering is lost, the same medications that once produced smooth benefit begin to produce wearing-OFF, delayed-ON responses, and dyskinesia. In advanced disease, oral therapy often becomes too short-acting and too burdensome to provide stable function, and the question becomes whether to move to device-aided or continuous-delivery therapy.

Across all stages, nonmotor disease increasingly determines disability. Cognitive impairment, psychosis, autonomic failure, sleep disturbance, constipation, urinary symptoms, depression, and anxiety often become more consequential than tremor or rigidity. A useful treatment framework must therefore address both the changing motor pharmacology and the growing nonmotor burden.

Initial Symptomatic Therapy: Levodopa Should Not Be Delayed for Its Own Sake

For many years, clinicians deferred levodopa out of concern that it might accelerate disease progression or exhaust its own effectiveness. That concern is no longer supported by the best available evidence. ELLDOPA (NEJM 2004) raised the possibility of imaging-detected harm when dopamine-transporter imaging suggested accelerated decline despite clinical improvement — a paradox now attributed to a pharmacologic imaging effect rather than neurotoxicity. The delayed-start LEAP trial (NEJM 2019) subsequently found no meaningful difference between early-start and delayed-start levodopa at 80 weeks (adjusted between-group difference in total UPDRS, −1.0; P=0.44), arguing against both neurotoxicity and disease modification. The practical conclusion is that levodopa should be started when symptoms warrant treatment, not withheld to preserve a theoretical future benefit.

The initial choice is usually among three broad approaches:

CALM-PD (JAMA 2000) captured the central trade-off between levodopa and dopamine agonists: initial pramipexole produced fewer dopaminergic motor complications (28% vs 51%) but less symptomatic benefit and more somnolence, edema, and hallucinations. PD MED later reinforced the practical value of levodopa-first treatment by showing better patient-rated mobility and quality of life than levodopa-sparing strategies over long-term follow-up. For most older adults, levodopa is the most rational starting point. Dopamine agonists remain reasonable in selected younger patients, but only after direct discussion of impulse-control disorders, daytime somnolence, and hallucinations. (Safinamide, often grouped with the MAO-B inhibitors, is approved in US practice only as an adjunct for OFF episodes, not as initial monotherapy, and is therefore discussed below.)

The hope that an early therapy might also be neuroprotective has repeatedly failed. DATATOP (NEJM 1993) showed that selegiline delayed the need for levodopa by roughly nine months, but the effect was symptomatic rather than disease-modifying, and vitamin E had no effect. ADAGIO (NEJM 2009) produced a delayed-start rasagiline signal at 1 mg that the 2 mg dose did not reproduce, and the inconsistency has prevented acceptance as proof of neuroprotection. At present, early treatment in Parkinson’s disease remains symptomatic treatment.

Motor Fluctuations and Dyskinesia: The Pharmacokinetic Era

As the disease advances, the therapeutic problem changes. The issue is no longer whether the patient responds to levodopa, but how steadily that response can be delivered. Wearing-OFF reflects underexposure late in the dosing cycle; peak-dose dyskinesia reflects overexposure at the peak; delayed-ON and dose failures reflect erratic absorption, often worsened by gastroparesis or food effects. Management becomes an exercise in smoothing dopaminergic delivery rather than simply increasing it.

The usual sequence is pragmatic: shorten the dosing interval or use a longer-acting levodopa formulation; add a COMT inhibitor or MAO-B inhibitor; add amantadine if dyskinesia is limiting; and provide an on-demand rescue for sudden OFF periods. When patients require increasingly frequent doses yet still spend meaningful portions of the day OFF, oral therapy is approaching its ceiling.

When Oral Therapy Is No Longer Enough

One of the most consequential errors in Parkinson’s care is referral for advanced therapy only after a patient has become too frail, cognitively impaired, or psychiatrically unstable to benefit fully. The commonly used 5-2-1 criterion is a useful screen: consider referral to advanced therapy when a patient requires five or more levodopa doses per day, has two or more hours of OFF time per day, or has one or more hours of troublesome dyskinesia per day. Any one of these should prompt a structured discussion.

EARLYSTIM (NEJM 2013) established that advanced therapy is not merely a salvage strategy for end-stage disease. In patients with relatively early motor complications (mean disease duration 7.5 years), subthalamic deep brain stimulation plus medical therapy improved quality of life by 8.0 points on the PDQ-39 compared with medical therapy alone, with better motor outcomes and fewer levodopa-induced complications. Referral should therefore occur while the patient is still a strong candidate, not after candidacy has been lost.

Advanced Therapies

Advanced therapy comprises three broad modalities — deep brain stimulation, MR-guided focused ultrasound, and continuous infusion. They are not interchangeable: they differ in reversibility, adjustability, effect on dyskinesia, reliance on hardware, psychiatric and cognitive tolerability, and suitability for unilateral versus bilateral symptoms.

Deep brain stimulation

Deep brain stimulation is the most established advanced therapy, with randomized evidence of superiority over best medical therapy for motor function, fluctuation control, and quality of life in appropriately selected, levodopa-responsive patients. The most relevant bedside decision is the target. CSP 468 (NEJM 2010) found that subthalamic (STN) and pallidal (GPi) stimulation produced similar motor improvement at 24 months, but STN allowed greater medication reduction at the cost of more decline in visuomotor processing speed and more worsening of depression; NSTAPS (Neurology 2016) found greater off-medication motor benefit with STN at three years, while GPi required more reoperations. In practice, STN is favored for the younger, cognitively intact patient in whom medication reduction is a goal, and GPi when dyskinesia is prominent or when mood or cognition is a concern, because the pallidal target is generally more forgiving neuropsychiatrically. The principal contraindications to STN-DBS are significant cognitive impairment, active psychiatric instability, and poor levodopa responsiveness — not age alone.

Adaptive (sensing-enabled) deep brain stimulation, which adjusts stimulation in response to recorded basal-ganglia activity, is best understood as an evolution of DBS programming rather than a separate therapeutic class. It is clinically relevant and likely to expand, but long-term comparative outcome data remain limited.

MR-guided focused ultrasound

MR-guided focused ultrasound creates a precise thermal lesion through the intact skull, with no incision, no hardware, and no implanted battery. In a sham-controlled trial, unilateral focused-ultrasound subthalamotomy improved the contralateral motor score by 8.1 points more than sham on MDS-UPDRS part III at four months, with the greatest benefit in asymmetric, tremor-dominant patients (NEJM 2020); dyskinesia occurred in nearly half of treated patients but was usually transient. Its defining limitations are that the lesion is irreversible and non-adjustable, and that bilateral lesioning carries higher risk to speech and gait, so it is largely a unilateral therapy. It is best suited to carefully selected patients with markedly asymmetric or tremor-dominant disease, reluctance toward implanted hardware, or limited candidacy for implanted therapy.

Continuous infusion therapies

Continuous infusion addresses the underlying pharmacokinetic problem directly: oral levodopa is intermittently absorbed and erratically delivered, whereas infusion provides a steadier dopaminergic signal. Continuous levodopa infusion is available in two forms — levodopa-carbidopa intestinal gel (LCIG), delivered into the jejunum through a PEG-J tube, and subcutaneous foscarbidopa-foslevodopa, which provides continuous levodopa without brain surgery or an abdominal tube but is associated with frequent infusion-site reactions and more limited long-term experience. Continuous subcutaneous apomorphine infusion is an effective non-surgical option for advanced fluctuations, particularly when OFF time is the dominant problem and the patient is not an ideal DBS candidate; nausea, orthostatic effects, infusion-site nodules, hallucinations, and impulse-control risk remain important limitations.

Choosing Among Advanced Therapies

A practical selection framework is built around a small set of patient axes:

• Symmetry and phenotype. Highly asymmetric or tremor-dominant disease may favor unilateral ultrasound lesioning; bilateral, fluctuation-dominant disease more often favors DBS or infusion therapy.
• Cognition and psychiatric status. Significant cognitive impairment argues against STN-DBS and should make any invasive therapy discussion more conservative.
• Dyskinesia burden. GPi-DBS or continuous levodopa delivery is often attractive when dyskinesia is the dominant problem.
• Reversibility and adjustability. When the long-term trajectory is uncertain, adjustable and reversible options are usually preferable to irreversible lesioning.
• Patient preference and logistics. Avoidance of implanted cranial hardware, a PEG-J tube, or a continuous pump are legitimate determinants of choice.

Recent Therapeutic Additions

The most useful recent advances have largely been refinements in delivery and adjunctive therapy rather than new disease biology. Extended-release levodopa formulations, including IPX-203, aim to lengthen Good ON time; opicapone offers clean once-daily COMT inhibition; safinamide is a useful adjunct for OFF periods; istradefylline provides a non-dopaminergic route to reduce OFF time; extended-release amantadine has improved the management of dyskinesia; and inhaled levodopa and subcutaneous apomorphine provide on-demand rescue for sudden OFF episodes. These are meaningful improvements in symptomatic management, but they do not alter the underlying neurodegenerative process. Several newer data sources — real-world experience with opicapone and IPX-203, and sensing-enabled (adaptive) DBS — are still maturing and should be regarded as emerging rather than established.

Disease Modification: Still Unresolved

The search for disease-modifying therapy continues, but no approach has entered routine practice on the basis of convincing evidence. Glucagon-like peptide-1 (GLP-1) receptor agonists remain investigational: a lixisenatide trial (LIXIPARK, NEJM 2024) suggested a modest signal on motor progression at 12 months, but the effect was small and accompanied by gastrointestinal toxicity, and the subsequent phase 3 exenatide program (Exenatide-PD3, Lancet 2025) did not establish a disease-modifying benefit. Anti-alpha-synuclein programs have likewise not yet produced a practice-changing therapy: prasinezumab did not demonstrate definitive slowing of progression in a phase 2 trial (PASADENA, NEJM 2022), and later signals remain exploratory rather than conclusive. The current standard of care therefore remains symptomatic and complication-directed.

Management of Motor Complications

Management of Nonmotor Complications

Many of the most disabling features of Parkinson’s disease are nonmotor, and they are routinely undertreated because they are under-asked. They also illustrate a recurring point: many are iatrogenic, driven by the dopaminergic and adjunctive drugs used to control the motor disease, so the first step is often to identify and remove offending agents — dopamine agonists, anticholinergics, amantadine, sedatives, and unnecessary polypharmacy — before adding new treatment.

Supplementary Table: Pharmacologic Therapy at a Glance

A consolidated reference to the major drug classes discussed above.

Conclusion

Treatment of Parkinson’s disease has become more flexible and more sophisticated, but it remains fundamentally symptomatic. Levodopa should be used when symptoms require it; motor complications should be approached as problems of delivery and timing; and advanced therapy should be considered while the patient remains a strong candidate rather than after candidacy has eroded. The major remaining gap is not symptomatic management but disease modification, where every modern program — MAO-B inhibitors, GLP-1 receptor agonists, and anti-alpha-synuclein antibodies — has so far failed to alter the disease trajectory. That gap continues to define the limits of current therapy and should be stated plainly.