The success rate was so striking that officials ended the study 14 months early to allow participants in the placebo arm to share in the benefits of the drug.

This has been demonstrated in seventeen randomised controlled trials, of which five included a placebo arm.

Compliance was 60% in the isotretinoin arm and 75% in the placebo arm.

Approximately 19% of patients had a prevalent vertebral fracture at baseline and the mean lumbar T-score of -3.0 in both active and placebo arm.

Five- and 10-year survival rates were 30% and 17% in the clodronate arm versus 21% and 9% in the placebo arm.

The 'no treatment' placebo arm was removed from the trial in June 2012.

A statistically significant RR of 43% for invasive breast cancer persisted at follow-up despite the addition of women to the placebo arm.

Neither total cancer nor breast cancer incidence was statistically significantly different between supplementation groups and the placebo arm.

Almost half (45%) of children in the placebo arms of these trials got better after taking the dummy pills, compared to 36% of adults.

Note that both arms of this study, the Bromday arm and the placebo arm, were not given any sort of corticosteroid to control inflammation.