Tuesday 19 May 2015

Placebo trials of surgery

In a recent systematic review of placebo trials of surgery (here) it was found that in half of the 53 trials found, surgery was not better than placebo treatment. And in the ones where it was better, the difference wasn’t great. This may not be big news to my readers, but this review was important because it highlighted many of the problems with surgery, namely that:
1) placebo studies are needed to determine the true effectiveness of surgical procedures, but …
2) surgery and associated devices are regulated less strictly than drugs, consequently …
3) surgery is often not subjected to placebo / sham studies, even though …
4) such studies are ethical and practical.

The reason for performing the review was partly because of the lack of appreciation and understanding of the placebo and non-specific effects associated with surgical procedures. The publication of this study will help that understanding.

There is disappointment in the fact that the studies included were mainly on minor procedures; over half were either inserting balloons (to stop bleeding) or using endoscopes. Many of the other, more significant procedures have been covered previously in my blog.

There is also disappointment in the response to the study, or rather the lack of it. Nobody cares much, and practice certainly hasn’t changed. In fact, one (of the very few) responses provided in the BMJ is from Russia, where it is lamented that so many outdated, ineffective and harmful surgical procedures (like radical mastectomy, vagotomy and gastrectomy for ulcers, and lung denervation for asthma) are still routinely performed in parts of the former Soviet Union.

The bottom line
The good news is that this review of placebo trials of surgery will open the doors for more placebo controlled trials of surgical procedures. There is a need for such trials, preferably before procedures become commonplace. This study goes some way to making such studies culturally and ethically acceptable, and by showing that such trials can be (and have been) done, it is more likely that researchers will consider using placebos in future studies.


  1. Dear collegue,
    I read since the beginning your blogs with high interest and agree with your conclusiones. In this study form Oxford , new Finnish study on meniscus surgery is missing. It would even strengthen the conclusions. A new meniscus study is underway by Stefan Lohmander from Lund in Sweden with the hyposthesis, that partial meniscectomy even leads to a higher and earlier onset of arthritis than non-operative treatment with physiotherapy etc.
    A very interesting study was published by Lohmander in 2013, showing that non-operative treated patients with ACL injury had after 5 years not more or even less arthritic changes than the operative reconstructed group Frobell RB, Roos HP, Roos EM, Roemer FW, Ranstam J, Lohmander LS: Treatment for acute anterior cruciate ligament tear: five year outcome of randomised trial. BMJ 2013; 346: f232).
    A few years ago we have established an Internet platform for second medical opinion (www.medexo.com). After having consulted around 1000 patients with orthopaedic problems, we could avoid more than 70% of the proposed surgeries. After 2 1/2 years these patients were still satisfied with their non-operative treatment. Interesting is that the non-operative treated group was 18 days earlier back to work than the operative treated group.
    I would be happy to exchange experiences with you in the future. Prof. Dr. Hans Paessler

    1. Thanks Hans. Unfortunately I couldn't work out what your organisation is all about because the website is in German. Is there an English language version of the site?

  2. Which surgical procedures (currently in use) would you most like to see placebo tested?
    Which do you suspect would not outperform sham, but would likely never get tested due to "ethics".

    Personally I'd like to see:
    - Tonsillectomy
    - THR
    - TKR
    - TSR (most shoulder procedures actually, especially "rotator cuff repair")
    - etc

    I'd be interested to know your gut feeling on a few surgeries currently inoculated from critisism by confirmation bias and the institution of medicine.

    1. Great question.

      First to cover your suggestions. Tonsillectomy, yes. THR and TKR yes, but difficult, as the rebalancing of alignment (mainly in Total Knee Replacement) is hard to control for. In some of the less severe patients (mild radiographic changes) this can be done and it wouldn't surprise me if knee replacement didn't come out much better than sham.
      "Most shoulder procedures" - yes. They are doing a sham surgery trial for shoulder decompression in the UK now. I am involved in designing an RCT for reverse shoulder replacement versus non-op treatment for bad shoudler fractures.

      Now, what other ones?
      Spine fusion for non-specific pain.
      Hernia repair.
      Excision of high grade brain tumours.
      Bariatric surgery (lap banding).
      Coronary stenting
      Most revascularisation procedures.
      Hysterectomy for dodgy indications.
      Floating kidney (yes, people are still doing this).
      Some fracture surgery.
      Achilles' tendon repairs.
      Spinal cord stimulators (almost certainly placebos).
      Almost any procedure you can think of for non-specific pain.

      And that's without even thinking about it too hard. Pick a specialty - they've all got their share of dodgy procedures.

  3. How about surgeries for femoral acetabular impingement?

    -LOVE your blog and have been a reader for years since I found your spinal surgery as placebo blog post way back when. Please keep up the great work!

    1. Thanks for the suggestion. FAI is a good one and I am currently doing some observational data linkage work on this, but there are no good comparative trials out there that I am aware of. I will keep a look out. In the meantime, surgeons will continue to increase their uptake of this unproven procedure before anybody gets any funny ideas about testing it.


Note: only a member of this blog may post a comment.