Prediction of drug photodegradation from stress testing by multivariate analysis

Nowadays, light has been recognized as one of the most important external factors in drug instability and the knowledge of drug photo-reactivity is essential for a correct pharmaceutical formulation and its packaging. The major effect from the degradation of most photolabile drugs is the loss of their pharmacological activity, but several cases involving the formation of toxic degradation products are known. Endogenous alterations caused by the light can also occur after drug administration for interaction of the drug with endogenous molecules resulting in a biological effect. Knowledge concerning drug photo-chemistry helps in the design of more photo-stable drugs as well as in the development of photo-protective systems. Photo-degradation studies should be performed by exposing the drugs to the direct solar light but the irradiation intensity is never constant and it changes significantly according to the seasons, the place and the atmospheric conditions. Furthermore, the exposure times necessary to obtain useful experimental data result to be rather long. The official rules for the photo-stability control of the drugs, by the institute ICH, establish stressfull light testing by diffused artificial light sources. The results from the photo-stability accelerated tests should be finally relationed with the virtual times under natural conditions necessary to cause the same degradation. The aim of the present study was the definition of relationships able to extrapolate the real degradation as a function of the degradation data from forced conditions. Investigation has been performed on the photo-sensitive drugs Nitrofurazone, Melatonin, Piroxicam, Amiloride and Lacidipine, submitted to ligth exposure under both sun and artificial light. Natural light radiation has been monitored along a three-year period in terms of illuminance (lux) and irradiation power (W/cm2). Since it resulted highly affected by the atmospheric conditions and day time, a reliable mean irradiation could be carried only from a so long monitoring period. Relationships between natural and artificial irradiation, based on PCR (Principal Component Regression) and PLS (Partial Least Squares) modelling, were established so that to extrapolate a real drug degradation from a forced degradation test.