Many pharmaceutical QC tests such as dissolution testing, content uniformity, assay, and blend uniformity require sample filtration prior to HPLC / UHPLC analysis. Since quantitation of analytes is critical following these tests, filter validation studies should be used to evaluate analyte loss due to membrane filter adsorption. Different membrane filters can bind analyte to varying degrees depending on membrane and analyte type, as well as analyte concentration.
The objective of this study was to provide guidance on filter selection during method development and validation with a special emphasis on analyte binding to syringe filters. The following membrane and analyte characteristics were evaluated as part of this study:
Drug dissolution studies were performed using multiple commercially available formulations and methods outlined in respective USP monographs. Samples were filtered using different syringe filters and various filtrate fractions were collected. Filtrate was analyzed by HPLC for quantitation of active pharmaceutical ingredient (API). Centrifuged samples were used as controls for 100% analyte recovery to calculate analyte binding to syringe filters. Table 1 lists formulations used in this study. Table 2 lists the dissolution and HPLC methods used.
Similar studies were also conducted on a blend uniformity sample provided by one of our customers. The sample was dissolved in a solvent blend and filtered through various syringe filters. HPLC analysis of the filtrate was carried out and recovery was calculated using a standard prepared in the same way.
Volumetric sample recovery from various syringe filters was determined by filtration of a fixed volume of sample through the syringe filter and measuring the volume of liquid collected in a vial. This provided information about volume retained by a syringe filter and its impact on analyte binding.
Analyte binding is mostly dependent on physico-chemical properties of both the membrane and analyte, since binding results from various secondary interactions between analyte and membrane. Some common secondary interactions that lead to analyte binding are electrostatic interactions, hydrogen bonding, and hydrophobic interactions.
Table 3 shows binding of different analytes from a multi-component migraine formulation using three different Millex® syringe filters containing hydrophilic PTFE, Durapore® PVDF, or nylon membrane. Active Pharmaceutical Ingredients (APIs) present in this formulation had different physico-chemical properties. This formulation contains an acidic (acetyl salicylic acid), neutral (caffeine), and basic (acetaminophen) API. As can be seen in Table 3, both acetaminophen and acetyl salicylic acid bind strongly to Millex® filters containing nylon membrane but not to Millex® filters containing hydrophilic PTFE or Durapore® PVDF membranes. On the other hand, caffeine doesn’t show binding to any of the three membranes selected in this study.
Both hydrophilic PTFE and Durapore® PVDF membranes have very few functional groups that can interact with various analytes, thereby leading to low analyte binding and subsequently high recovery. Nylon membrane contains amino and carboxylic acid functional groups as well as amide bonds which can interact with acidic or basic analytes through electrostatic and hydrogen bonding interactions, leading to high analyte binding and low recovery. Analyte recovery with nylon syringe filters can be improved by simply saturating the syringe filter with the sample, thereby saturating the binding sites present on the membrane. This can be clearly seen in Table 3 for second, third, or fifth 1 mL of filtrate.
Quantitative analyte recovery can be obtained after the filter is fully saturated with the analyte. When the membrane filter is saturated, no further analyte binding is observed because syringe filters have limited surface area as well as a limited number of functional sites that enable analyte binding. This saturation point is dependent on analyte concentration. Generally, as the analyte concentration decreases the volume needed to fully saturate the filter increases.
Figure 1 shows the effect of analyte concentration on analyte binding and subsequent recovery of naproxen from a formulation. Three different concentrations of naproxen ranging from 244 ppm to 2.4 ppm were filtered through hydrophilic PTFE Millex® syringe filters and filtrate was analyzed by HPLC. As can be seen in Figure 1, no analyte binding was observed for the high concentration of naproxen (244 ppm), even without any filter saturation (no discard volume). At lower concentrations, slightly lower recovery was obtained for first 1 mL of filtrate due to incomplete saturation of the syringe filter but quantitative recovery can be obtained for third or fifth mL of filtrate (indicating filter saturation). Since hydrophilic PTFE membrane generally shows low analyte binding, this concentration effect was subtle, but still can be significant if the recovery specification was very narrow.
Figure 1.Effect of analyte concentration on analyte recovery. Naproxen dissolved in phosphate buffer (pH 7.4) was filtered through a 0.45 µm hydrophilic PTFE Millex® syringe filter.
This study was also conducted with 0.2 µm Millex® hydrophilic PTFE syringe filters, with very similar results obtained (Figure 2). Membrane pore size did not significantly impact naproxen saturation of syringe filters. This concentration effect is membrane- and analyte-dependent and hence during filter validation studies, discard volume should be selected carefully depending on analyte concentration.
Figure 2.Effect of analyte concentration on analyte recovery. Naproxen dissolved in phosphate buffer (pH 7.4) was filtered through a 0.2 µm hydrophilic PTFE Millex® syringe filter.
Processing conditions of standard and sample can impact analyte binding and subsequent analyte recovery. Good lab practices suggest that the sample and standard should undergo the same processing to avoid any impact of processing conditions on analyte recovery.
Analyte recovery using blend uniformity samples provided by our customer was evaluated. Standard was initially prepared using a pure compound dissolved in the solvent mixture, whereas the blend uniformity sample was dissolved in the same solvent mixture and then filtered using a Millex® Durapore® PVDF syringe filter. Five mL of filtrate volume was discarded before sample was collected for HPLC analysis. Our customer observed that lower than expected recovery was obtained indicating incomplete saturation of syringe filter (data not shown).
We repeated this study wherein sample was processed as per the original protocol, but the standard was processed using three different conditions:
Results are shown in Table 4. When Standard 2 was used for recovery calculations, consistent and quantitative recovery was obtained for all the syringe filters tested. This is expected since both the standard and sample undergo the same processing method, thereby reducing any impact of analyte binding on recovery.
On the other hand, when Standard 1 or 3 was selected for recovery calculations, a negative impact was observed on sample recovery due to analyte binding. This effect was more pronounced with Millex® Durapore® PVDF syringe filters compared to Hydrophilic PTFE Millex® syringe filter. Some lot-to-lot variability was also observed for using the Durapore® PVDF syringe filter.
In most cases, pharmaceutical QC tests are not sample-limited and volumetric sample recovery may not largely impact sample availability. However, when calculating analyte binding, volumetric sample recovery plays an important role since the volume required for filter saturation needs to account for the volumetric retention of syringe filters.
Figure 3 shows sample volume recovered when 2 mL water was filtered through various 25 mm and 33 mm syringe filters. Of all the filters tested, Millex® filters (33 mm diameter) allowed for maximum sample volume to be recovered (~1.4 mL) when filtering 2 mL water, whereas the polypropylene membrane-based syringe filter (25 mm diameter) only recovered 0.6 mL of sample volume. This volume retention is dependent on the filter design and is not largely impacted by membrane pore size. With the other three membrane filters tested, approximately 1 mL of sample was retained within the syringe filter.
Figure 3.Volume recovery after filtration of 2 mL water through various syringe filters. The hydrophilic PTFE Millex® filters are 33 mm in size, whereas all the other syringe filters (from vendor B, C, D and E) were 25 mm in size. Membranes tested include hydrophilic PTFE, regenerated cellulose (RC), and polypropylene (PP).