The European Medicines Agency (EMA) has opened a new consultation on a draft guideline that would revise a set of biosimilar guidelines first put in place six years ago.

The guideline, On non-clinical and clinical development of similar biological medicinal products containing recombinant human insulin and insulin analogues, was first proposed in March 2005, and later finalized in February 2006. In November 2012, EMA's Biosimilar Medicinal Products Working Party (BMWP) released a draft revision of the policy, which was adopted by the Committee for Medicinal products for Human Use (CHMP) on 13 December and subsequently released to the public.

Some sponsors of biosimilar human insulin products, including Marvel LifeSciences Ltd, have recently withdrawn their marketing authorization applications (MAAs) in light of the planned revisions, saying they need more time to repeat and resubmit bioequivalence, pharmacokinetic and pharmacodynamics data using the new standards established in the guideline.

The Guideline

EMA's guideline notes that non-clinical studies on biosimilar insulin should be designed to "detect differences in the response to the similar biological medicinal product and the reference medicinal product and should not just assess the response per se." Those studies should include pharmacodynamics studies-both in vivo and in vitro-to assess differences in how a drug is broken down once introduced into the body and, in certain cases involving novel excipients, toxicological studies. Generally, though, "separate repeated dose toxicity studies are not required," EMA explained.

The guideline devotes a significant amount of time to clinical studies, and in particular pharmacology studies-what EMA calls "the mainstay of proof"-- showing a biosimilar product is equivalent to a reference product in safety and efficacy.

EMA calls for a homogenous, insulin-sensitive study population afflicted with type-1 diabetes-but otherwise healthy-to be studied in a pharmacological trial. Studies may include either just men or both men and women. EMA explains that insulin sensitivity may "vary during the menstrual cycle," and that this could potentially affect study results. "The inclusion of only men in the studies would be justified," it explained.

The Clamp Method

The majority of the guideline, however, goes on to explain the finer details of the "clamp" method, in which the plasma insulin concentration of a patient is raised using the selected therapy while the blood-glucose level (GIR) of the patient is maintained, or "clamped," at its current level through the injection of glucose. Measuring both the concentration and the infusion rate of the glucose allows a sponsor to estimate the time-concentration and time-action profile of the therapy, EMA explained. Sponsors may study the time-concentration and time-action-profiles of the products separately, though EMA says it prefers if they are studied simultaneously.

"It is not easy to control the blood glucose concentrations during the clamp study," EMA explained. "The Applicant should provide an estimate of the quality of the performance of the clamp study, e.g. by calculating the coefficient of variation of the blood glucose concentrations. The mean intra-individual coefficient of variation of well executed euglycaemic  hyperinsulinaemic clamps should usually not exceed 10% for glucose infusion rate."

Other Considerations

These considerations will be different for long-acting insulin products, EMA said, and will likely require a much more difficult process involving clamping in order to prove biosimilarity. "Taken together, hyperinsulinaemic euglcycaemic/isoglycaemic insulin clamps, with some limitations, may be appropriate to compare the time-concentration and time-action profiles of long-acting biosimilar and reference insulins but will usually require a relatively large sample size and a long duration for the purpose of demonstrating similarity," EMA wrote.

Pharmacokinetic data, meanwhile, should show the biosimilar product to be within the 80% to 125% AUC and C_max endpoints at the 90% confidence interval-the conventional acceptance range for bioequivalence. Sponsors may make the case for an alternate measurement if they are able to justify the rationale for such a change. Other endpoints include the T_max, T_50%, and T_1/2, though all are secondary endpoints.

Comments on the guideline are due by 30 June 2013.