Issue link: https://www.e-digitaleditions.com/i/1276475
12 August 2020 Inhalation since either their ISM or their MMAD are outside the assigned limits. Each APSD is also provided in one of the two right-hand panels according to its classification (either within specifications or out-of-specification (OOS)), allowing readers to visually evaluate the suit- ability of the proposed specification limits. Comparing the red OOS APSDs in the lower plot to those falling within the specification limits (upper plot, in blue), one can easily rationalize the rejection of each OOS APSD. Each such distribution appears to deviate from the typical performance of the product in either one or both dimensions. In other words, the proposed MMAD and ISM limits appear as though they would perform as intended; they would allow the user to make sensible batch disposition decisions. Let's recap briefly: we have taken 201 APSDs rep- resenting the performance of an inhaled product, and assigned realistic limits to both the mass and size dimensions of the APSD, resulting in the sensible rejection of those APSDs whose mass or size deviates from the product's typical performance. Although one can always debate the specific values of these limits, we contend that these are reasonable, realistic specifica- tions and that they do the job they are intended to do. Specifically, these specifications would allow your QC laboratory to make sensible batch disposition decisions by discriminating between batches with APSD profiles Although the actual specifications for this product are not known, we can still assign reasonable lim- its based on the spread of the data. To control the MMAD, we choose here to set upper and lower lim- its at ± 0.5 microns from the median MMAD value of 1.53 microns (at 2.03 microns and 1.03 microns, respectively). To control the ISM, we choose here to set upper and lower limits at ± 33% of the median ISM value of 54.2% label claim (LC) (at 72.1% LC and 36.3% LC, respectively). ese ranges are cho- sen based on the authors' experience developing and characterizing OIPs in an attempt to define reasonable limits for this data set, allowing for realistic variability (inter- and intra-batch variability, inherent method variability, etc.). It is worth noting that the ± 33% range applied to the ISM would encompass approxi- mately one standard deviation of a normally distrib- uted population, and is consistent with specification limits of approved products. 7 Laying the limits for each metric (shown by the four labeled reference lines) atop the plot of ISM vs. MMAD in Figure 2 divides the plot into nine regions in a 3 x 3 grid. e central region defines those APSDs (shown in blue) that are considered acceptable by these specifications; both their ISM and MMAD are within the assigned upper and lower limits. APSDs falling in any of the eight peripheral regions (red data) are deemed unacceptable by these specifications, Figure 2 Control of the product's APSD by imposing limits on both the size and mass dimensions. The main plot expresses each APSD in terms of ISM vs. MMAD, with the labeled reference lines defining the proposed upper and lower limits for each metric. Blue and red symbols indicate whether APSDs fall within or out-of-specification, respectively. At right, the 201 APSDs are divided between two plots, allowing visual inspection of the proposed specifications' classification of APSDs as either within (top, in blue) or out-of-specification (bottom, red). 80 70 60 50 40 30 ISM (%LC) MMAD (microns) Median MMAD ± 0.5 microns Within Specification Out-of-Specification (OOS) Median ISM ± 33% 36.3 1.03 2.03 1.0 1.5 2.0 2.5 3.0 3.5 35 30 25 20 15 10 5 0 APSDs within Specification 0 1 2 3 4 5 6 7 8 9 APSDs Out-of-Specification 35 30 25 20 15 10 5 0 Recovery (%LC) Recovery (%LC) 0 1 2 3 4 5 6 7 8 9 Aerodynamic Diameter (microns) Aerodynamic Diameter (microns) 72.1