Bridging the efficacy–effectiveness gap

Bridging the efficacy–effectiveness gap: a regulator's perspective on addressing variability of drug response
Hans-Georg Eichler, Eric Abadie, Alasdair Breckenridge, Bruno Flamion, Lars L. Gustafsson, Hubert Leufkens, Malcolm Rowland, Christian K. Schneider & Brigitte Bloechl-Daum

Abstract

Drug regulatory agencies should ensure that the benefits of drugs outweigh their risks, but licensed medicines sometimes do not perform as expected in everyday clinical practice. Failure may relate to lower than anticipated efficacy or a higher than anticipated incidence or severity of adverse effects. Here we show that the problem of benefit–risk is to a considerable degree a problem of variability in drug response. We describe biological and behavioural sources of variability and how these contribute to the long-known efficacy–effectiveness gap. In this context, efficacy describes how a drug performs under conditions of clinical trials, whereas effectiveness describes how it performs under conditions of everyday clinical practice. We argue that a broad range of pre- and post-licensing technologies will need to be harnessed to bridge the efficacy–effectiveness gap. Successful approaches will not be limited to the current notion of pharmacogenomics-based personalized medicines, but will also entail the wider use of electronic health-care tools to improve drug prescribing and patient adherence.

Nature Reviews Drug Discovery 10, 495-506 (July 2011)

Looking at PK data #2

Here is a parallel set of data comparing young healthy adults and patients with cirrhosis. Again, this set of data is not meant to prove changes which happen in cirrhosis but more as an exercise to explore and understand PK effects.

As in the previous post, the first thing that strikes you is the massive changes in AUC seen in cirrhotics. The AUC for cirrhotics is 1.85 times that of the young. By the same logic as in the previous post this AUC increase is inversely related to a decrease in clearance and/or bioavailability. Since there is no way to assess bioavailability in this set of data, we shall leave it out of the discussion for the time being. It is not that the cirrhosis did not result in a change of bioavailability, it is just that the experimental design does not allow us to examine bioavailability effects.

Unlike the situation with the elderly, the reduction in clearance in cirrhotics is not accompanied by a fall in the unbound fraction. Instead, the unbound fraction is elevated from 3.5 to 5.3%. Changes in protein binding is not uncommon in liver cirrhosis. Often serum albumin is reduced. Here the protein binding is decreased with a corresponding increase in the free fraction. In fact the increase in free fraction has the effect of actually increasing metabolic clearance.The reduction in clearance is therefore not a result of a decrease free fraction, but primarily due to loss of enzyme activity. One can in addition, expect that the extent of degradation of enzyme activity is even greater than indicated by the extent of decrease of clearance because part of this effect is mitigated by an increased clearance caused by the increase in free fraction.

The increase in elimination halflife and corresponding fall in Kel is related to the reduction in clearance. The magnitude of the change in halflife is however larger than the fall in clearance and suggests that perhaps the Vd may have increased as well. This is expected because of the increase in free fraction.

Looking at PK data #1

Here is a set of data, adapted from published information comparing pk data of orally administered Drug X between young subjects and elderly subjects. The data here just provides an opportunity to qualitatively discuss PK effects, and it is not the intention here to 'prove' any PK changes in the elderly.

In this case the most apparent difference is rather large and significant difference in AUC between young and elderly. The AUC in elderly is about 1.84 times greater than that for the young. This is not an unexpected finding. The question is what is the cause of the reduced AUC?

We know from theory that AUC is determined primarily by bioavailability and clearance. However, since we have no way to assess bioavailability here, we shall concentrate on clearance effects instead. The increase in AUC is consistent with a reduction in clearance in the elderly. Again from theory, assuming this is primarily metabolic clearance, we expect that metabolic clearance of an orally administered drug is dependent on protein binding and enzyme activity.

When we inspect the protein binding data we find that the protein binding in elderly is actually increased with the unbound fraction falling from 4.3 to 3.4%. But this magnitude of change is relatively small compared to the estimated change in clearance. Hence it is possible the the total change in clearance reflects both a reduction in unbound fraction as well as a degradation of enzyme activity.

The fall in the unbound fraction potentially also affects the Vd. We have no direct way of assessing the Vd changes here but the Cmax provides indirect (though inaccurate) look at possible Vd changes. The Cmax for the elderly is higher than in the young but marginally less (probably insignificantly less) than the magnitude of change for AUC, so the Vd effect is uncertain.

The halflife changes in the elderly are consistent with the reduction in metabolic clearance.

The uncertainty in this case study is how much any bioavailability changes play in affecting the PK data. The magnitude of halflife change is quite comparable to the magnitude of AUC and clearance changes. This suggests that if there are bioavailability effects it is probably minimal.