University of Chicago Journal Club, March 2011
Editor, B Roitberg, MD
Williams A, Gill S, Varma T, Jenkinson C, Quinn N, Mitchell R, Scott R, Ives N, Rick C, Daniels J, Patel S, Wheatley K; PD SURG Collaborative Group. Deep brain stimulation plus best medical therapy versus best medical therapy alone for advanced Parkinson's disease (PD SURG trial): a randomised, open-label trial. Lancet Neurol. 2010 Jun;9(6):581-91.
The PD SURG trial we are discussing today is an open label randomized ‘intent to treat’ trial comparing medical treatment alone with combined medical and surgical treatment. The trial was conducted across 13 neurosurgical centers in the UK from 2000 to 2006. The investigators intended to prove the usefulness of deep brain stimulation in treating Parkinson’s patients by showing a 10-point difference between both groups on PDQ 39 (Parkinson’s disease Questionnaire). Secondary endpoints included changes in UPDRS (Unified Parkinsons Disease Rating Scale) and DRS II (Dementia Rating Scale II) at the end of 1 year following treatment. This article presents the follow up data at 1 year.
The plan was to enroll about 400 patients to generate a statistical significance with p<0.005 for 90% confidence intervals. To be eligible to participate patients needed to have a confirmed diagnosis of Parkinson’s disease, DRS II scores greater than 5 and be fit for surgery. The study was structured to create randomization through computer minimization using various categories including age, time since diagnosis, Hoehn and Yahr Stage, Indications for surgery, type of surgery and drugs used in medical treatment. Pair-wise randomization in individual institutions aimed to balance some of the treatment variation across centers. The study did not specify the exact type of medical or surgical therapy – they choice was left to the discretion of the local clinician. Comparison between different medical therapy agents was made by converting all medications into levodopa equivalents. Neuropsychological assessment was planned as part of the study but could not be performed for all the participants. Complications and adverse effects of surgery and medication respectively were collected.
In 6 years a total of 366 patients were enrolled, divided equally between both groups. Most of the participants (348/366) were less than 70 years old and 341/366 subjects had Parkinson’s for more than 5 years. Dyskinesia and severe off periods were the typical reasons for considering surgery. In studies that perform an intent to treat analysis it is important to minimize crossovers between groups so as to maintain the clinical validity of the study. Only 5 individuals in the surgical arm were unable to get surgery and 12 patients in the medical group received surgical treatment before 1 year.
Surgery appeared to provide superior results. On PDQ-39 summary index (The PDQ-39 is the most widely used Parkinson's Disease specific measure of health status. It contains thirty nine questions, covering eight aspects of quality of life. University of Oxford Department of Public Health), the investigators found a significant mean difference of -4.7 in favor of the surgical group. ADL (activities of daily living) showed the most improvement followed by mobility and bodily discomfort. Social stigmata associated with the disease were least affected. A significant difference (mean -8.3) was noted on UPDRS. The benefit of surgery, as expected, quadrupled when subjects were tested during ‘Off periods’. At 1 year significantly more patients in the surgical group were free from waking day dyskinesia and reported no off periods. Verbal fluency and executive function testing showed a small decrease in the surgical group in comparison with the medical group on the neuropsychological tests. Surgical patients at year were at a 34% lower levodopa equivalent dose which was statistically significant too. Adverse events and complications were higher in the surgical group as expected.
In summary, the PD SURG trial was well constructed and executed. The limitations of the study include being an open label study involving varying clinical practices especially with medical treatment of the disease. Very minimal crossovers and almost perfect follow up rate give results additional weight. The study is not original - the results mirror most other similar studies. One serious disadvantage is the lack of specifics of care. This leaves a lot of unanswered questions.
There was remarkably low crossover. Maybe the rate of crossover would have been higher in the US. In Britain this study was the only way to get deep brain stimulation for PD. Those who wanted to get it but were not randomized to it had to pay a hefty sum out of pocket; if I am not mistaken something around 100 thousand dollars. It is also interesting to note that essentially all the surgical procedures were subthalamic DBS.
I think the lack of change in social stigmata described in the study means that even after surgery patients felt no increase in social acceptance.
It is not entirely clear what specific scientific questions were asked. The benefits of DBS were demonstrated years earlier. After reading this study it is difficult to have a good understanding of the risk or benefit of any particular treatment modality, since the study design lumped all medical treatment into one group, and all surgical treatments into another. It so happened that the surgical interventions were quite uniform, but the medical were not.
Maybe the main question was economic – should we pay for surgery? A study so large must have cost a fortune, yet saved the British government money because as long as it lasted they did not have to pay for every patient who would benefit from DBS.
Why was a trial of DBS compared to medical treatment published in 2010, after DBS was approved in 2002 by the FDA, and many other randomized controlled prospective studies showed that DBS is superior to medical treatment alone?
This study is too late and provides too little new information.. It was funded by the MRC (Medical Research Council) who also paid for the treatment costs in the NHS. The slow and restrictive British health care system also made this study possible because it limited crossover and guaranteed adherence to the protocol as well as long enough duration to get the needed numbers.
The limitations of the study are also evident in the hints about suboptimal medical care. There was a large difference between patient functional scores on DBS and on medication – greater than in other studies; this indicates that medical regimens were suboptimal. There was a smaller decrease in L-dopa equivalents after surgical treatment than in other studies – 34% vs typical 50%. And also the improvement in UPDRS was small compared to other studies. When the study reports surgical complications such as intracerebral hemorrhage, it does not mention if the same patients had microelectrode recording. It makes hypothetical sense that adding insertion of sharp electrodes may increase bleeding risk, but we cannot learn it from this study - a missed opportunity.
The decrease in cognitive function in the surgical group is also interesting – some cognitive side effects are common and maybe a price of DBS. This could have been studied better, and should be evaluated in future studies.
This brings us to the question of optimal targeting of DBS for PD – and the next article.
Follett KA, Weaver FM, Stern M, Hur K, Harris CL, Luo P, Marks WJ Jr, Rothlind J, Sagher O, Moy C, Pahwa R, Burchiel K, Hogarth P, Lai EC, Duda JE, Holloway K, Samii A, Horn S, Bronstein JM, Stoner G, Starr PA, Simpson R, Baltuch G, De Salles A, Huang GD, Reda DJ; CSP 468 Study Group. Pallidal versus subthalamic deep-brain stimulation for Parkinson's disease. N Engl J Med. 2010 Jun 3;362(22):2077-91.
The goal of this study was to compare alterations in motor function in patients with advanced Parkinson’s disease with placement of DBS electrodes in the subthalamic nucleus (STN) versus the globus pallidus (GPi) over the course of 24 months following surgery.
This double-blind, randomized trial with intention-to-treat analyses assigned 299 Veterans Affairs hospital patients to receive either STN or pallidal stimulation. Motor function was assessed over the course of 24 months at 3 month intervals using the Unified Parkinson’s Disease Rating Scale, part III (UPDRS-III); at each session, antiparkisonian medications were held overnight and the stimulator was turned off for one hour, then a dose of medication was given and the stimulator was turned back on. Secondary outcomes were also assessed, including changes in the Beck Depression Inventory score, use of dopaminergic medications, neurocognitive testing, adverse outcomes, and patient diaries of subjective motor function. Patients were included in the study if they were 21 years of age or older, stage 2 or higher on the Hoehn and Yahr disability scale off of medication, responded to levodopa, had persistent and disabling motor symptoms after pharmacological optimization, had at least 3 hours of poor symptom control in every 24 hour period, and had been on medical therapy without changes in agents or dosing for a month. The study was designed to detect a 25% difference in the change in UPDRS-III score between the groups at the 24-month evaluation with a power of 80%, with a P-value of 0.05 indicating statistical significance. This primary outcome was analyzed using a two-sample t-test, with additional mixed-effects model analysis to compensate for data missing at random. Patients with incomplete data had their last visit treated as the 24-month visit for evaluating changes after the procedure.
Briefly, the comparison between target groups demonstrated no statistical difference in UPDRS-III motor scores, adverse effects, or quality of life; increase in depression scores, decrease in processing speed index, decrease in dopaminergic medication dosage and lower average stimulation amplitudes in STN vs GPi targeting. The authors mention a previous study that showed the same primary outcome as the current study, but provided data only 6 months postoperatively. They also discuss a study that found increased quality of life with pallidal vs STN stimulation, which may reflect alterations in neurocognition and depression found in the current study. Finally, they reflect that the ability to reduce dopaminergic medication use with STN targeting may provide benefits of its own, and the reduced stimulation amplitude may allow patients to undergo battery changes less frequently.
The difference in test scores reached statistical significance, but it is small. Is the difference clinically important?
PD looks like depression to an observer. Is the lack of motion itself part of the depressive symptoms? Is there a biochemical change related directly to the disease, or is the depression a less specific reaction to illness?
Please note that the difference in the study is between averages. Individual patients could be affected less or more. This means that some patients who received STN stimulation may have crossed from mild to severe depressive symptoms. So I think the difference in depression scores was clinically significant for some of the patients.
The study demonstrates further that the choice of target is not a trivial matter, and that multiple cognitive and behavioral effects need to be studied and taken into account. However, are we sure what the targets are, anatomically? Is the target the STN or the white matter next to it?
Indeed, “STN stimulation” is something of a misnomer; our electrodes stimulate a 1 cm volume that includes the zona incerta, which is the probable real target.
This study was very well done. The design takes pharmakokinetics into account by standardizing all evaluations at definite timepoints after medication or cessation thereof. Patients could not cross over – they did not know what procedure was done because randomization was done in the operating room rather than in advance and patients were blinded to the stimulation site.
The difference in outcomes between the groups provides a rich source of information for tailoring the treatment for patients. Some of the differences are due to mechanical considerations. The STN is small, so microthalamotomy effects are more likely than micropallidotomy effects in the case of the GP. Other differences are due to other factors in the circuitry. STN stimulation is better for dyskinesia, especially since you want to decrease the L-DOPA dose. On the other hand, pallidal stimulation patients fared better on depression scores and verbal processing score of the WAIS. A particular pitfall is the risk of worsening depression as STN patients get less L-dopa. Removing L-dopa can have a depressant effect by itself.
Both cognitive decline and depression are prevalent in PD. They have to be taken into account when planning treatment. Even a few points decline on the depression scale can result in a notable clinical change. The results of this important study should be used by clinicians when counseling patients about their surgical options.