Placebo surgery #2: Surgery for Parkinson’s disease

In Parkinson’s disease, there is insufficient dopamine in parts of the brain. This chemical can be replaced in pill form, but with variable effect, and with some side effects. In the 1990’s many clinics were transplanting dopamine-producing cells from embryos into the brains of people with Parkinson’s disease. This procedure involved drilling holes in the skull, through which the cells were inserted. Animal studies showed that the cells could survive, and that the procedure could correct some of the movement disorders.  Open label trials (no placebo) of these transplants in humans showed that patients improved. This is the same evidence base for many operations performed today: a biological mechanism, supportive lab studies, and reports of patients that got better. Enter, the sham.

I had read about these trials, but had not read the articles themselves. When I was searching for them, I found more articles written about the ethics of these trials than actual trials published. The highest profile articles were firstly from the authors of one of the trials defending their use of sham surgery (here), and the following article against its use (here). After some searching, I found the actual trial that they were talking about (here) from 2003, and an earlier trial from 2001(here).

The trials

In the 2001 trial from the New England Journal of Medicine 20 patients were randomised to each group (fetal cells and sham surgery). The results for the primary outcome (the Global Rating Scale) were very similar between the two groups, with basically no improvement in either group. This is interesting, because my previous post on sham angina surgery showed that both groups improved, but by a similar amount. Now, given that they found no difference between the groups in their primary outcome, even when dividing them into young and old, why does the conclusion in their summary state: “Human embryonic dopamine-neuron transplants … result in some clinical benefit in younger but not in older patients”?

Because amongst their numerous sub group analyses of different outcomes for different genders and different age groups, they found (wait for it) that for some outcomes, in younger patients, when tested in the morning, before their medication, the transplant group did better than the sham group. Since when does a subgroup analysis like that trump the primary outcome on which the study (and the hypothesis) is based? And while they were keen to include all the p values (significance tests) for each subgroup, they didn’t do any significance testing on the finding that major adverse events occurred in 42% of the transplant group compared to 5% in the sham group. I did, and the p value (probability of arising by chance) for that is 0.006.

The quality of life outcomes for the same patients (reported in a separate study here), showed no significant difference between the surgery and the sham groups, and here both groups did improve. Interestingly, patients who thought they got the active treatment did significantly better than those who thought they got the sham, regardless of what they actually received (and yes, their guesses were no better than chance, which shows that the study was well blinded). If just thinking they got the treatment made them better, how much improvement would there be if patients knew they got the treatment (like when we report case series)?

The 2003 trial of 31 patients in three groups was published in the Annals of Neurology. Except for a few minor differences (general anaesthetic instead of local, and the addition of immunosuppression) they basically tested the same thing, right down to the sham surgery. The active group did a little better than the sham group, but the difference was not statistically significant. Inevitably, they found a sub group where the transplant fared better than the sham (ironically, it was the mild cases) but at least they had the scientific merit to correctly conclude that: “Fetal nigral transplantation currently cannot be recommended as a therapy for Parkinson’s disease based on these results”.

Questioning the biological plausibility (mechanism)

Like a lot of these things, they sound good superficially, but when you think about them for a while, they don’t. To me, putting dopamine producing cells into the brains of people with Parkinson’s makes as much sense as injecting bone cells into people with osteoporosis. It is similar to stem cell therapy, where the attraction lies in the theory, not the results (see previous blog: Stem cell therapy: still science fiction). To think that we can correct the vast complexities involved in a disease that we do not fully understand by injecting some cells from an embryo is a probably little naïve.