Journal Club · Functional Neurosurgery

STN vs GPi DBS for Parkinson Disease

Two targets, equivalent motor benefit, and a decision that turns entirely on the trade-offs.

A trainee reading of the randomized trials that define target selection in Parkinson disease — the VA CSP 468 program (Weaver and Follett) and the Dutch NSTAPS study — and what they actually license us to say about choosing the subthalamic nucleus or the globus pallidus internus.

Weaver JAMA 2009 Follett NEJM 2010 NSTAPS Lancet Neurol 2013 COMPARE Ann Neurol 2009

Why These Trials

Once a patient with advanced Parkinson disease has cleared candidacy for deep brain stimulation, the next decision is the target: the subthalamic nucleus (STN) or the globus pallidus internus (GPi). For two decades this has been one of the most-debated choices in functional neurosurgery, and it remains a near-certain board and clinic question. The evidence is unusually strong — several blinded randomized trials address it directly — and the honest summary is that the two targets are motorically equivalent, so the decision turns on a set of trade-offs in medication reduction, mood, cognition, dyskinesia, and programming.

One-Line Takeaway

Bilateral STN and GPi stimulation produce equivalent, durable motor benefit; choose the target to fit the patient’s vulnerability profile — STN for medication reduction, GPi for neuropsychiatric safety and direct dyskinesia control — not to chase extra motor gain, because there is none to be had.

Part I

VA CSP 468: The Definitive Head-to-Head

1.Stage One — DBS vs Best Medical Therapy (Weaver, JAMA 2009)

A randomized controlled trial across 7 Veterans Affairs and 6 university hospitals enrolled 255 patients with advanced PD (Hoehn & Yahr stage ≥2 off medication); notably, 25% were ≥70 years old, answering the common criticism that DBS trials exclude older patients. Patients were randomized to bilateral DBS (STN n=60 or GPi n=61) or best medical therapy (n=134). The primary outcome was time spent in the “on” state without troublesome dyskinesia, captured by motor diaries.

The DBS group gained a mean of 4.6 hours/day of good “on” time versus 0 for best medical therapy (between-group difference 4.5 h/d, 95% CI 3.7–5.4; P<0.001). 71% of DBS patients achieved a clinically meaningful (≥5-point) UPDRS-III improvement versus 32% with medical therapy, and quality of life improved on 7 of 8 PDQ-39 subscales. The cost was surgical: more serious adverse events in the DBS arm, and small decrements in some domains of information processing.

Why It Matters Weaver established the premise — bilateral DBS clearly outperforms best medical therapy for motor fluctuations in advanced PD, including in carefully selected older patients — and set up the target comparison that followed.

2.Stage Two — STN vs GPi (Follett, NEJM 2010)

Within the same program, 299 patients were randomized to bilateral pallidal (GPi, n=152) or subthalamic (STN, n=147) stimulation, with blinded UPDRS-III assessment. The primary outcome was change in UPDRS-III motor score from baseline to 24 months, on stimulation and off medication.

Motor outcome — no significant difference. Mean UPDRS-III improvement was 11.8 points (GPi) vs 10.7 points (STN); P=0.50. Self-reported function and quality of life also did not differ.
Medication — STN advantage. STN patients reduced dopaminergic medication significantly more than GPi patients (P=0.02). This is STN’s signature benefit.
Cognition — small STN cost. A measure of visuomotor processing speed declined more after STN than GPi (P=0.03).
Mood — GPi advantage. Depression (Beck Depression Inventory) worsened after STN and improved after GPi (P=0.02).
Safety. Serious adverse events were somewhat more frequent with STN, with no significant between-group difference overall at 24 months.

Long-term follow-up (Weaver et al., Neurology 2012) found motor benefit similar between targets and stable to 36 months.

Limitations A predominantly male, older VA population limits generalizability to younger and female patients; blinding to target is imperfect; and programming and medication management were not fully standardized across a large multisite network.

Part II

NSTAPS: Testing the “GPi Is Safer” Hypothesis

3.The Dutch Randomized Trial (Odekerken, Lancet Neurology 2013)

NSTAPS set out to test a specific hypothesis: that GPi would produce greater overall functional improvement than STN because of fewer cognitive, mood, and behavioral complications, with similar motor benefit. Five centers randomized 128 patients (65 GPi, 63 STN) 1:1, with masked assessors. There were two co-primary outcomes: functional health (the weighted Academic Medical Center Linear Disability Scale, ALDS) and a composite of cognitive, mood, and behavioral effects at 1 year.

Neither primary outcome differed. Weighted ALDS change was 3.0 (SD 14.5) for GPi vs 7.7 (23.2) for STN (P=0.28); the neuropsychiatric composite was essentially identical — adverse effects in 36/62 (58%) GPi vs 35/63 (56%) STN (P=0.94). The predicted GPi advantage did not materialize.
Secondary outcomes favored STN: greater off-drug motor (UPDRS-III) improvement and significantly greater medication reduction (mean levodopa-equivalent reduction 546 [SD 561] mg STN vs 208 [521] mg GPi; P=0.01).

Three-year follow-up (Odekerken et al., Neurology 2016) reinforced this: STN gave greater off-drug motor improvement and medication reduction, with no robust between-group difference in cognition or mood.

4.A Discussion Point — The COMPARE Trial

Worth raising: this single-center, randomized, blinded comparison of unilateral STN vs GPi (Okun et al., Ann Neurol 2009; n=45 completing) found no difference in mood or global cognition at optimized settings, but a greater decline in letter (phonemic) verbal fluency with STN that persisted with stimulation off — implicating the surgical trajectory rather than the current — and showed that stimulating ventral to the optimal contact worsened mood in both targets. The complementary message: much of the apparent “STN is bad for mood/cognition” effect is positional and programmable. (Detail from the published abstract; confirm against full text.)

Part III

Choosing The Target

5.The Trade-offs At A Glance

Domain	STN	GPi
Motor efficacy	Equivalent	Equivalent
Medication reduction	Greater (its main advantage)	Modest
Mood / depression	Can worsen (CSP 468); neutral in NSTAPS	More favorable / can improve
Cognition	Small decrements (fluency, processing speed)	Better preserved
Dyskinesia	Indirect (via medication reduction)	Direct antidyskinetic effect
Programming / energy	Narrower window, lower energy, longer battery	Wider window, higher energy use

6.Safety Caveats Worth Stating Out Loud

Mood, Apathy, and Suicide Risk Depression must be identified and controlled preoperatively, with a psychiatrist engaged, because of the recognized risk of post-DBS mood worsening and suicide — a particular concern after STN. The dramatic, early medication reduction that STN permits can itself precipitate apathy and dopaminergic-withdrawal phenomena; tapering should be deliberate.

Speech and Impulse-Control Disorders STN stimulation tends to improve speech less than other parkinsonian signs, and hypophonia or dysarthria may worsen from current spread to corticobulbar fibers. In patients with impulse-control disorders, medication weaning is sequenced thoughtfully (dopamine agonists first); STN in the setting of dopamine-dysregulation/impulse-control disorders has shown inconsistent behavioral benefit and can worsen it.

Questions For The Table

Is this patient’s leading problem medication burden (favor STN) or cognitive/mood fragility and dyskinesia (favor GPi)?
Has depression and impulse-control behavior been formally assessed and stabilized preoperatively?
How much does battery longevity and programming simplicity weigh for this patient and family?
Do the trial populations (older, male, VA) resemble the patient in front of you?
If motor benefit is equivalent, what is the single factor that should decide the target here?

Motor efficacy is equivalent and durable (24–36 months) across CSP 468 and NSTAPS — target choice is not about motor benefit.
STN’s reproducible edge is greater medication reduction (and lower energy use); GPi’s is more favorable mood and a direct antidyskinetic effect.
NSTAPS failed to confirm a clear GPi neuropsychiatric advantage — the gap is real but smaller and more selection-/programming-dependent than once believed.
Lean GPi for the older, cognitively or psychiatrically fragile, or dyskinesia-dominant patient; lean STN when medication reduction is the priority and cognition is intact.
Screen and stabilize mood and impulse-control behavior before surgery, whichever target is chosen.

Selected References

Weaver FM, Follett K, Stern M, et al. Bilateral deep brain stimulation vs best medical therapy for patients with advanced Parkinson disease: a randomized controlled trial. JAMA. 2009;301(1):63–73.
Follett KA, Weaver FM, Stern M, et al. Pallidal versus subthalamic deep-brain stimulation for Parkinson’s disease. N Engl J Med. 2010;362(22):2077–2091.
Weaver FM, Follett KA, Stern M, et al. Randomized trial of deep brain stimulation for Parkinson disease: thirty-six-month outcomes. Neurology. 2012;79(1):55–65.
Odekerken VJJ, van Laar T, Staal MJ, et al. Subthalamic nucleus versus globus pallidus bilateral deep brain stimulation for advanced Parkinson’s disease (NSTAPS study): a randomised controlled trial. Lancet Neurol. 2013;12(1):37–44.
Odekerken VJJ, Boel JA, Schmand BA, et al. GPi vs STN deep brain stimulation for Parkinson disease: three-year follow-up. Neurology. 2016;86(8):755–761.
Okun MS, Fernandez HH, Wu SS, et al. Cognition and mood in Parkinson’s disease in STN versus GPi deep brain stimulation: the COMPARE trial. Ann Neurol. 2009;65(5):586–595.