Karin Liltorp, Ph.D., is currently working as principal scientist at Particle Analytical Aps., an analytical laboratory performing physical chemical characterization of pharmaceutical drugs.
Most pharmacists have heard the frightening old story about Ritonavir: a drug against HIV that suddenly changed crystal structure, leading to a drastic reduction in bioavailability with severe consequences for the patients. The company (Abbott Laboratories) lost an estimated $250 million in sales as well as hundreds of millions of dollars in an attempt to recover the original polymorph/crystal form. This case made everybody aware of the existence of polymorphs and the danger of not having these under control.
Still, most solid pharmaceuticals are dosed as crystals. Almost all pharmaceuticals can arrange in different crystal lattices/polymorphs giving rise to different properties of the drug. As these polymorphs are not equally stable, a potential risk of a spontaneous transformation from a less stable to a more stable crystal form always exists.
The ICH guidelines state that a “polymorph screening,” i.e., a search for other crystal forms, should always be performed to ensure control of the crystal form in development and thereby control of bioavailability and stability. However, nothing is stated about the extent of the screening. Only imagination sets the limits for possible experimental procedures! The nature of crystallization is really fascinating, but unfortunately it is also quite unpredictable: You never know how many forms you will find, and you can never be sure that you have actually found the most stable form.
Some wise person once said that if you find only one polymorphic form, you did not search well enough. There are always different possibilities of ways for the molecules to attach; who knows if a new form will crystallize from a mixture of acetone and hexane at 43 degrees when adding water gently under stirring? Or maybe a form will crystallize when freezing a supersaturated water/ethanol solution fast in liquid nitrogen! It is not hard to picture that this screening can be quite time consuming and expensive. Herein lies a major challenge: understanding the difference between nice-to-have and need-to-have in order to properly balance the level of information needed and the development costs.
The experimental work put into a screening should be based on a risk assessment. What is the risk that other polymorphs exist (probably high)? What is the risk that they will have fundamentally different properties (depends on the molecule in question)? What impact will it have on bioavailability (depends on solubility and therapeutic window)? The more critical a transformation is, the more experimental work should be put into it. In order to create as many crystal forms as possible, you have to create as many different environments as possible. For instance, you will increase the possibility of a completely different conformation if you change the solvent properties maximally (e.g., water versus hexane). Further, creating a wide range of concentration might induce new ways of arranging. Overall, the experimental work does not have to be that extensive—as long as the experimental parameters are based on a scientifically well founded approach. The earlier you perform the screening the better, as this will de-risk your development program and potentially save you unpleasant and expensive surprises later on (like the Ritonavir case).