MD, PhD, FMedSci, FRSB, FRCP, FRCPEd.

If you thought that osteopathy is for spinal problems, think again. This study aimed to determine whether osteopathic manipulation of the T9-T10 vertebrae improves the evolution of tonsillitis. A randomized, stratified, controlled clinical trial with blinded patients, evaluator, and data analyst was performed.

The patients in the control group (CG) received a “sham” manipulation consisting of a 150° passive flexion of the shoulders, with a gentle contact of the osteopath’s knees in the middle thoracic vertebrae, without impulse or causing tension. A high-speed, low-amplitude technique was applied to the T9-T10 vertebrae in the osteopathic manipulative group (OMG) patients.

The number of days needed to resolve the tonsillitis was significantly lower (p = 0.025) in the OMG (2.03 ± 0.95 days) than the CG (2.39 ± 0.82 days). Additionally, the number of episodes of tonsillitis after the treatment decreased significantly more in the OMG (0.8 ± 1.88 episodes/year in total) than the CG (2 ± 2.12) (p = 0.005). In the OMG, 60.8% had no recurrences of tonsillitis, compared to 22.5% of the CG, in the following year (χ2 (1) = 15.57, p < 0.001). No patients reported adverse effects.

The authors concluded that during an episode of tonsillitis, the number of days to resolution was significantly lower after the application of an osteopathic manipulation of the T9-T10 vertebrae, compared to a sham manipulation. The number of subsequent year tonsillitis episodes was greatly reduced in both groups, significantly more in the OMG than in the CG patients.

This is an interesting study. Its exceptional feature is that it is sham-controlled. This must mean that the results are reliable and that osteopathic manipulation is indeed an effective treatment of tonsillitis.

Or perhaps not?

It is, of course, most laudable to introduce a sham control group into such a study. But let us for a moment reflect on what purpose it serves. Its main purpose, no doubt, is to render patients unaware of which treatment they received, sham or verum. This is only possible if the sham is indistinguishable from the verum. In this study, this was clearly not the case. Patients in the verum group felt that they received the verum, and patients in the sham group felt that they had the placebo. Thus the former expect to get better, while the latter don’t.

So, we are left with two very different interpretations of the findings:

  1. Osteopathic manipulations are effective for tonsillitis.
  2. Osteopathic manipulations are not effective for tonsillitis, but the patients’ expectations determined the outcomes.

Which is more likely?

What do you think?

3 Responses to Osteopathic Manipulations for Tonsillitis?

  • So, the difference in duration between the two groups was 8 hours 38 minutes 24 seconds (with a tolerance of the best part of a day), determined by once-a-day phonecalls,..

    • I imagine the researchers standing by will a stopwatch, ready to press the call button on their phone at precisely the right time…

      Any calculation can be reported to an arbitrary level of precision, but as you point out, that is usually quite meaningless, and precision is anyway completely different from accuracy. Di the authors not have access to somebody with some basic knowledge of mathematics to help them analyse their data?

      By the way, I’m not sure that tolerance is strictly the same as confidence intervals. I have not come across the term tolerance in medical statistics; I understand it as being more akin to precision and used in engineering in the context of measurements and manufacturing.

      The number of days needed to resolve the tonsillitis was significantly lower (p = 0.025) in the OMG (2.03 ± 0.95 days) than the CG (2.39 ± 0.82 days).

      This statement (taken from the abstract of the paper) confirms that the authors really don’t understand their statistics. Here they express the difference between the two groups both in terms of statistical significance and as confidence intervals, giving contradictory results. Confidence intervals are generally preferred as being more representative of the data, and here there is very substantial overlap, meaning that it is unlikely that there is any true difference in outcome between the two groups.

      Statistical analysis was carried out with SPSS 22.0 software

      I have taken this quote from the original paper, which is free to view here:
      https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8065872/
      SPSS is widely used in medical statistics. However, it doesn’t make you a statistician any more than Word makes you a writer or Quick Books an accountant.

  • If only they’d manipulated the T8/9 vertebrae!
    More research is clearly called for.
    Sigh.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Recent Comments

Note that comments can be edited for up to five minutes after they are first submitted but you must tick the box: “Save my name, email, and website in this browser for the next time I comment.”

The most recent comments from all posts can be seen here.

Archives
Categories