Issue link: https://www.e-digitaleditions.com/i/1229406
Inhalation April 2020 21 recent publication by our group demonstrates just how poorly suited FPD is to the QC environment, using APSD data from four different OIPs. 4 As in the exam- ple here, the authors observed that FPD is remarkably insensitive to APSD size changes. Building on this published work, the two graphs in Figure 4 summarize individual APSD data from eight OIPs. e first graph plots FPD values versus impactor-sized mass (ISM). e second graph plots the same FPD values versus their cor- responding MMADs. Note that each data point is a sin- gle measurement from an individual inhaler. e pMDI featured in the prior example is included as Product A. Two observations are immediately obvious from these graphs: FPD correlates very strongly with ISM, and not at all with MMAD. Each of these observations is true for each individual OIP and also for the combined data set. e strong correlation between FPD and ISM shown in the left graph arises from the fact that the FPD con- stitutes the vast majority of the ISM. e dependence of ISM on FPD is, in fact, so great that they cannot be used as separate metrics. In other words, using both met- rics (ISM and FPD) would offer no improvement in QC decision-making compared to the use of either metric alone. Meanwhile, the lack of correlation between FPD and MMAD, laid bare in the right graph, confirms that, for these eight OIPs, the vast majority of their size-frac- tionated mass is finer than the 5 µm FPD boundary. is may seem like a simple restatement of the previous point (that FPD and ISM are essentially one and the same), but it reinforces an important message: that FPD does not reliably detect shifts in MMAD. Even for Product E (an HFA suspension MDI) whose MMAD is closest to the 5 µm boundary, shifts in MMAD from roughly 3.5 µm to 4.4 µm fail to impact the FPD. by an MMAD in excess of 3 µm. This represents a dramatic and readily observable departure from the 1-2 µm MMAD range typically observed for this prod- uct. However, this large shift in the size dimension of the APSD has resulted in only a slight decrease in FPD from > 95% to roughly 85%. e small drop in FPD is clearly just the tip of the iceberg. In short, FPD is remarkably insensitive to changes in the aerodynamic size of the aerosol. Context and implications So what does this mean? Some may doubt the signif- icance of this example, questioning, in particular, the clinical relevance of shifts in MMAD. If a 1 µm shift in MMAD is deemed to have no clinical significance, why should we care if such shifts are detected by FPD? is argument completely misses the point of QC testing. e disparity between the current batch's APSD and those of previous batches tells us that something has gone wrong, that something in either the formulation, the process or the device components is not as it should be. Whatever the cause of this unexpected outlier APSD, it reflects that the product or the process is not under control and the root cause(s) should be deter- mined. is is true regardless of whether there are clear implications regarding efficacy and/or safety for the patient. is concept is fundamental to quality control, within the pharmaceutical industry and beyond. Stepping back, it is important to appreciate what the preceding example represents. It is natural to question whether the phenomenon illustrated here is a common occurrence or whether this data set represents an isolated peculiarity. The key lesson from this example—the insensitivity of FPD to changes in the size dimension of APSDs—is not limited to a specific OIP. Indeed, a Figure 4 FPD for eight different OIPs plotted against ISM (left panel) and MMAD (right panel). Each data point represents an individual APSD determination from the IPAC-RS blinded database (see sidebar for details). 80 70 60 50 40 30 20 10 0 FPD (%LC) 0 10 20 30 40 50 60 70 80 90 ISM (%LC) Product A: HFA Solution MDI B: DPI C: CFC Suspension MDI D: CFC Suspension MDI E: HFA Suspension MDI F: CFC Suspension MDI G: DPI 80 70 60 50 40 30 20 10 0 FPD (%LC) 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 MMAD (microns)