Abstract
Purpose :
The globule size (GS) and size distribution of ophthalmic emulsions (OEs) are two critical quality attributes of the drug product. Measuring GS in the nanometer (nm) range was challenging by Mie light scattering due to strong Rayleigh scattering of nm GS; the OEs were “invisible” even at high concentration (Conc.). Therefore, a DLS method, which is suitable for GS in nm ranges was developed to meet the critical need.
Methods :
A DLS method for determination of GS distribution of OEs was developed referencing ISO 22412:2008 (Particle Size Analysis--DLS). Per this ISO, two robust descriptors (Z-average or Zv and polydispersity index or PdI) are recommended for measuring the GS of an emulsion. One pre-requisite for reporting the mean size by Zv is that the sample scattering distribution peak must be Gaussian, without peak separation. Also, according to this ISO, if the sample’s apparent Zv and PdI are conc. dependent, either the values extrapolated to infinite dilution or those obtained at a near zero conc. or C0 should be used.
Results :
The conc. effect on apparent Zv and PdI was tested by diluting neat samples 2 to 300-fold with water. The Zv of the OEs was linearly dependent on conc. with least square coefficients (R2) > 0.99 shown in Figure 1. OE Conc. vs apparent Zv or App Z-ave). Zv of 31 nm obtained by extrapolating the apparent Zv to C0.
For qualitty control purpose, it is impractical to perform multi-conc. analysis of each OE on routine basis. Therefore, there is a need to select an OE conc. that is low enough to yield an apparent Zv close to that obtained by extrapolating to C0, with acceptable precision. Based on Zv test results of five lots of the OEs, at a conc. after 50-fold dilution, each apparent Zv was very close to that determined by obtaining Zv vs. conc. plot and extrapolating to C0. The results showed < 5% RSD (n=6).
Conclusions :
A DLS method was developed to measure GS distribution of OEs using the approach described in ISO 22412:2008. The apparent Zv of the ophthalmic emulsion is dependent on sample conc. and the Zv value may be obtained by extrapolating to C0. However, we demonstrated that measuring the Zv and Pdl with an appropriately diluted sample, without having to construct a Zv vs. conc. plot, could also provide equivalent and precise results (<5% RSD) for QC testing purpose. This was demonstrated by using product lots of various size distributions.
This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.