In 2008, the US Food and Drug Administration (FDA) released a Guidance for industry on evaluation of cardiovascular safety of new drugs for type 2 diabetes [1]. This promoted large-scale global cardiovascular outcome trials (CVOT’s), where the addition of the investigational drug to existing treatment was compared to placebo. So far, three trials on dipeptidyl peptidase 4 (DPP-4) inhibitors and four trials on glucagon-like peptide 1 (GLP-1) receptor agonists have been reported, all in the New England Journal of Medicine [2,3,4,5,6,7,8]. In total, the staggering number of exactly 70,000 patients were randomized in these trials. Overall, the results range from reassuring to positive, but also give rise to new uncertainties. These uncertainties stem from the much larger datasets on efficacy and safety now available for these new drugs, thus providing us with much more granular knowledge than we used to have for drugs on the market. In other words, with these datasets also the frequency of more infrequent side effects can be assessed with some detail. However, even these large-scale trials are of course not powered for really rare side effects, adding a new level of uncertainty if a non-significant signal is seen.
Overall, the DPP-4 outcome trials confirmed the cardiovascular safety of this drug class, although the safety of their use in patients with known heart failure is not established. A higher rate of hospitalization for heart failure was seen with saxagliptin [2]. A similar signal emerged for alogliptin, although limited to the subgroup without a history of heart failure [3]. There was no heart failure signal in the sitagliptin CVOT [4]. Possible mechanistic explanations for DPP-4 inhibitor related heart failure are emerging [9].
Another approach to scrutinize the available datasets is to subject them to meta-analysis. Pancreatitis was numerically more frequent in the active treatment arm in each of the three DPP-4 outcome trials, but the signal didn’t reach significance in any of the individual trials. Taking these three numerical imbalances together in meta-analysis, pancreatitis could be classified as a rare side effect of DPP-4 inhibitors [10].
GLP-1 analogs: more than safety alone
The picture for GLP-1 analogs is more diverse, and perhaps more exciting, as two, or perhaps even three of the four published CVOT’s reported positive rather than neutral outcomes. The first trial to report- on lixisenatide- was neutral. This is probably related to the pharmacokinetics of lixisenatide, because the drug is not active for 24 hrs when given twice daily as per label. The other three GLP-1 agonists with reported outcome trials have a duration of action that covers the full 24 hours. The liraglutide outcome trial showed a meaningful reduction in the primary outcome, 3-point MACE, as well as in cardiovascular and total mortality [6]. This was confirmed for 3-point MACE in the much smaller cardiovascular safety trial for semaglutide [7]. The exenatide once-weekly CVOT just missed significance for superiority in the primary outcome (p=0.06 for 3-point MACE), precluding further statistical analysis because a stopping criterion in the hierarchical statistical testing sequence was met [11]. However, the relative risk and 95% confidence interval for total mortality (RR 0.86 (0.77-0.97)) would indicate agreement with the benefits seen for liraglutide and semaglutide. There is an ongoing discussion on the underlying mechanisms, including the question whether the combination of improved glycemic control, blood pressure reduction and body weight seen with these drugs fully explains their cardiovascular benefit. In addition, two trials with liraglutide in patients with heart failure, although not powered for clinical outcomes, showed worse cardiovascular outcomes in the GLP-1 receptor agonist arm [12, 13].
Again using the meta-analysis tool, pancreatitis could be refuted as a side effect of the GLP-1 class. And also the discussion on pancreatic cancer as a possible side effect of this class could be put to rest [14].
In conclusion, large-scale CVOT’s, despite being costly, have provided us with a wealth of information. Whereas Profil is focusing on the first part of a drugs’ lifecycle, rather than on the last part, it is important to be aware of the full spectrum when embarking on the development of a new drug for diabetes.