Placebo surgery #1: Surgery for angina

I will attempt to write a short note on every clinical study that has compared surgery to a placebo operation (sham surgery). The common thread is that while many patients improve after the real operation, many patients also improve after the sham operation, and in each example the real operation is no more effective than the sham surgery. This shows that when we see patients get better after we treat them, it is not necessarily due to the specific effect of our treatment. We (the doctors and the patients) perceive effectiveness by attributing cause and effect to the association we see. The difference between the perceived effect and the specific (real) effect is the placebo effect.

This study dates from 1959, and relates to a surgical procedure for angina that dates from 20 years before that: internal mammary artery ligation.

The procedure involves blocking (ligating, or “tying-off”) an artery that runs inside the chest, near the heart, in order to divert more blood to the heart. The operation had everything going for it: biological plausibility, support from animal experiments, and good results from series of patients that had the procedure. Already, you have about as much evidence as many of the surgical procedures performed today. And this was not considered an experimental procedure; it was performed in many institutions across the USA.

The 1959 sham surgery experiment (link) involved surgery on all of the participants, but half of them didn’t actually have the artery ligated, and the patients were ‘blinded’ to the type of treatment they received. Most patients in both groups felt better (in fact all patients in the study had improved exercise tolerance), but they were no more likely to improve if they had the surgery than if they had the sham surgery, showing that the treatment did not have a specific therapeutic effect, only a perceived effect. Basically, the operation didn’t work. The study was small (17 enrolled) which might make you think that it was not a powerful enough study to detect the advantage of surgery, but I can counter that with the observation that the patients in the placebo group actually did better than the (real) surgical group for most outcomes. A larger study may have been more conclusive about a difference between the groups: it might have been able to show us that the surgery was significantly harmful.

The good news is that the operation fell out of favour after publication of this study. Similar results from trials of knee arthroscopy for arthritis (link) have not resulted in a similar decrease in rates of surgery (definitely not in my country, anyway). It is amazing how long an operation can ‘linger’ after it has been shown to be ineffective (see my blog on blood letting here). We need to rely more on what the science tells us than on our gut feelings.


And if you think that surgery is much better for angina nowadays, I direct you to a previous blog post (here).

Placebo surgery #2: surgery for Parkinson's disease is here.