Due to Camargo’s on-going client projects in this area, our chief medical officer, Dr. Sam Kaba recently attended a DIA-sponsored conference in Washington, D.C. entitled Cardiovascular Safety and Development of Type 2 Diabetes Mellitus Medications: Current State of the Art and Opportunities to Advance the Science. His report:
The FDA Guidance for Industry on Evaluating Cardiovascular Risk in New Antidiabetic Therapies to treat Type 2 Diabetes came out in December 2008. Since then pharmaceutical and biotechnology companies involved in diabetic drug development have been trying to put their arms around the new requirements included in the guidance, and understand their implications to their development programs. The full impact of this guidance goes way beyond just adding a risk evaluation and mitigation plan; it rather introduces a whole new paradigm for developing new drugs for type 2 diabetes.
This guidance has a sound scientific rationale: CV events are common in type 2 diabetes patients, so how can we be sure that the risk of CV events is not increased by a new antidiabetic drug? The answer is by calculating the hazard ratio of vascular events in patients treated with the new agent compared to the control group. Sounds simple…but it is not!
First, the Guidance currently limits the CV events that can be included in this analysis to non-fatal myocardial infarcts, non-fatal strokes, and death from CV origin. These are referred to collectively as Major Cardiovascular Adverse Events (MACE). The annual rate of MACE in type 2 diabetes patients is around 2%.
Guidance sets the upper bound of the 95% CI for the hazard ratio of MACE to 1.3 to allow the approval of the new product without post marketing cardiovascular safety trials. This upper bound was determined after multiple discussions with internal experts in other divisions in the FDA and external experts. This limit was also used as a cutoff in assessing the CV risk of other products such as COX2 inhibitors. To meet this upper bound a very large number of patients have to be enrolled in clinical trials. For example, if both treatment groups have the same annual MACE rate of 2%, 8,000 patients will be needed. While if the annual event rate in the control group is 1.75 and in the new drug group is 2% the number jumps to 34,000. Clearly, if the drug under study is CV protective, i.e. the hazard ratio is < 1, the number of patients needed will be substantially lower.
Since these patient numbers are not practical for pre-approval trials, the Guidance goes a step further to allow the initial approval of new products when the upper bound of the 95% CI is less than 1.8, with the condition that the sponsor produces post approval data to confirm that the upper bound is actually less than 1.3. Even with an upper bound of 1.8, around 1,500 patients will need to be treated for at least 1 year in controlled trials. This represents a significant increase the number of patients and the duration of trials for type 2 diabetes. Diabetes studies typically enrolled around 300-400 patients and lasted 24 months.
To meet these numbers of patients and MACE events, the sponsor, has the option of:
- Combining data from multiple late phase trials and conducting a meta-analysis of the MACE events,
- Conducting large pivotal trials with efficacy endpoints at 6 months and safety endpoints at 1 year, or
- Conducting a separate pre-approval CV safety study that runs parallel to the efficacy trials.
Additionally, if the upper bound of the 95% CI detected during the trials is > 1.3, the sponsor still has to conduct an additional post marketing CV safety study to confirm that the upper bound is actually less than 1.3.
According this new Guidance, the sponsor is expected to submit a CV Risk Management Plan that details the planned trials, the means of collecting the CV events data, the type of CV data collected (MACE only or other events), and the details of the statistical plans including the meta-analysis, if applicable. This CV Risk Management plan should be submitted prior to starting trials in diabetic patients (i.e. Phase II).
Although this Guidance is “final” the FDA acknowledge that it is work in progress. There are ongoing discussions regarding what vascular events to include, the optimal design of trials, the details of the meta-analysis, etc. These discussions may yield some decrease in the number of patients needed, if events other than MACE are allowed to be included in the risk ratio analysis. The FDA is planning to ultimately merge this CV risk guidance with the final version of the Diabetes Guidance that was issued in a draft form in February 2008.
It is not clear at this point if these rules will apply to 505(b)(2) programs referencing older antidiabetic agents that were approved prior to Dec. 2008. What is clear, however, is that the upper bound rules will not apply to insulin products. As with any guidance from the FDA, this is only a broad outline of the clinical program that does not replace the need for direct and open discussion with the Agency on the specifics of a new product.