IONSYS™ versus morphine PCA: Analysis of the current literature using a Bayesian approach

Abstract

Background

The IONSYS™ device, providing transdermal fentanyl, has been found as non-inferior to intravenous morphine PCA in four non-inferiority studies; the accepted loss of efficacy had to be lower than a 10% decrease of therapeutic success. Bayesian analysis allows to compute likelihoods pertaining to any chosen difference in therapeutic rate of success, and is not limited to conclusions based on arbitrary thresholds.

Methods

We try to show the superiority of intravenous morphine PCA, using the data of the four comparative studies. We have subjected the four studies to a Bayesian analysis, with respect to the common primary outcome reported in the four studies, in a sequence that parallels their dates of publication. In a first analysis we used an uninformative prior, and in a second analysis a moderately sceptical prior.

Results

With an uninformative prior probability, there is a 75% probability that some (>0%) increased rate of success exists with morphine PCA. With a moderately sceptical prior the probability is still around 69%. The probability of a relative risk reduction of treatment failure >10% of difference, is 14.8% with an initial uninformative prior and 3.7% with a sceptical prior. The probability of a relative risk reduction >20% is virtually non-existent.

Conclusions

Although, some decrease in the rate of therapeutic success can be expected, the fentanyl transdermal system seems a viable substitute to intravenous morphine PCA, as a large difference in the rate of therapeutic success is unlikely.