Extractables and leachables

Method validation means full validation as per ICH Q2 guidelines.

Methods for targeted leachables studies need to be validated, i.e. the method is validated for a specific compound in a specified matrix.

Methods for extractables screening studies need to be qualified.

Qualification involves the application of a subset of the parameters listed in the ICH Q2 guidelines to a model compound/standard. The method is then applied to a variety of compounds in a variety of matrixes.

SST is used to verify that the chromatographic system is adequate for the intended analysis.

For specified standard(s):

• Precision – repeatability
• Accuracy – spike recovery
• Limit of detection (LOD).

It is sufficient to include one matrix in the method qualification, except for the LOD/limit of quantification (LOQ), which should be determined in all four extraction solvents (50% Ethanol, 0.5 N NaOH, 0.1 M H₃PO₄, water). It is recommended to use 50% ethanol as a matrix in the qualification of gas chromatography-mass spectrometry (GC-MS) and liquid chromatography mass spectrometry (LC-MS), and water in the qualification of headspace gas chromatography mass spectrometry (HS-GC-MS).

For specified standard(s):

• Precision – Replicate injections (n=6)
• Sensitivity, e.g. by running a standard of concentration close to LOQ
• Retention time.

SST standards are prepared in a solvent suitable for the analysis.

Two separate SST samples may be prepared: one spiked at a higher concentration (e.g. 1ppm) to assess specificity and precision, and one spiked at a concentration close to LOQ to assess sensitivity. Alternatively, additional compound(s) can be spiked at a low concentration to the high-spike SST sample to assess sensitivity. Another option is to use a high-concentration spike SST, e.g. 1ppm, and set a relevant signal-to-noise (S/N) ratio requirement to assess sensitivity.

Retention time can be a good indicator of system performance. No pass criteria are given for retention time. Noting the retention time of the recommended standards also facilitates a comparison between laboratories of chromatographic range

HS-GC-MS:

• Toluene
• Methylethyl ketone (MEK)
• Octamethylcyclotetrasiloxane
• Toluene-d8 (internal standard).

GC-MS:

• N-Octane (and/or Eicosane)
• Butylated hydroxytoluene(BHT)
• Phenanthrene-d10 (internal standard).

LC-UV:

• Bisphenol A (BPA)
• Irganox™ 1010

LC-MS:

• A relevant standard of your choice to verify MS sensitivity.

Note: In the Best practices guide for evaluating leachables risk from polymeric single-use systems in biompharmaceutical manufacturing, a modified set of recommended standards was published, compared to Standardized extractables testing protocol for single-use systems in biomanufacturing published in 2014 (BioPhorum protocol)Methanol was replaced by Toluene and n-Octane was replaced by Eicosane. This has been implemented in the revised recommendations from BioPhorum on extractables testing of biomanufacturing equipment.

The SST requirements for LC-UV-MS are specified for ultraviolet (UV) detection. It is, however, recommended to use relevant compounds of your own choice to verify the MS performance, as Bisphenol A and Irganox 1010 may not ionize efficiently in all four mass spectrometry detection modes.

The BioPhorum protocol includes recommended standards to facilitate the comparison between laboratories. Sensitivity is evaluated by comparing the LOD of the recommended standards. The chromatographic range is evaluated by comparing the relative retention times of the recommended standards.

The method qualification can be documented in the methods tab in the BioPhorum Extractables Data Summary (BEDS) spreadsheet template, in a separate section in the extractables test report or a standalone document. Chromatograms of the SST standard should be made available on request. Link to BEDS

It is not mandatory to use the recommended standards and conditions. However, use of the recommendations facilitates comparison between laboratories.

We have included recommended standards to demonstrate similar chromatographic range using different columns and instruments. The widest chromatographic range is desired. That is also the reason for specifying the standards. In addition, the relative retention compared to known standards can be assessed for identified (even tentatively) compounds.

We do not want to be too rigid since many laboratories have an acceptable methodology, but we want to include enough detail to ensure acceptable results are obtained.

The following two approaches are commonly used for quantitation of extractables:

• • Approach 1: An external authentic compound is used in a one-point calibration curve or a multi-point calibration curve. If an authentic compound is not available, a surrogate compound can be used
  • Approach 2: An internal standard is added to the sample. Relative response factors are then used for the quantitation of the extractable compounds. The relative response versus the internal standard is determined for each extractable compound in a separate experiment by using an authentic compound of each compound. If an authentic compound is not available, a surrogate compound of a similar structure can be used. If the structure of the extractable is unknown, an assumption can be made that the response of the unknown extractable is identical to the response of the internal standard, i.e. the relative response factor is assumed to be 1.

The following approaches are commonly used for selecting peaks for reporting:

• Approach 1: The total ion chromatogram or UV signal of the sample and the control are compared. The total ion chromatogram is commonly processed, e.g. into a base peak chromatogram before comparison
• Approach 2: An algorithm is applied that automatically assigns peaks from individual/groups of ion traces in both samples and controls, which are then compared.

An approach where the samples are only screened versus a pre-set list of target compounds is generally not acceptable.

The following approaches are commonly used for setting a peak reporting threshold:

• Approach 1: Peaks fulfilling the area criteria of ≥3 times the control and equal to or higher than the area of an external standard of specified concentration
• Approach 2: Peaks fulfilling the area criteria of ≥3 times the control and ≥Y% of an internal standard of specified concentration.

Y is selected based on the concentration and response of the internal standard and the analytical evaluation threshold (AET). The AET is related to the drug dosing, which is unknown. A general AET of 0.1µg/mL (ppm) is thus used in the BioPhorum protocol. This corresponds to a drug intake of ~10mL/day, which covers many drug products.

It is recommended to use the nomenclature in the United States Pharmacopeia (USP) chapter Assessment of extractables associated with pharmaceutical packaging/delivery systems, which designate individual extractables identifications in the categories of Confirmed, Confident or Tentative. It is also recommended to add the category Unknown.

Each compound needs to be reported by it Chemical Abstracts Services (CAS) registry number, if available. It is recommended to use a descriptive chemical name. It does not have to conform to Union of Pure and Applied Chemistry (IUPAC) nomenclature. Only using trade names is not recommended. To help the reader, trade names, etc. can be added in brackets after the chemical name.

It is recommended to use a nomenclature of Unknown 1, Unknown 2, etc. to be able to distinguish between different unknowns. In the Excel spreadsheet, each unknown will be associated with a method, retention time and semi-quantitative result. Also, the highest masses need to be denoted. It is not mandatory to include chromatograms and spectra in the report. These should, however, be recorded and made available upon request.

The list of elements is based on USP and ICH Q3D. Also, elements that have been shown to potentially affect drug quality are included (Aluminum, Iron, Magnesium, and Zinc). It is optional to include additional elements.

It is only required to record the results from the final extractables testing time point. It is optional to analyze samples from additional time points.

ICP-MS analysis is only required for water, 0.1M H₃PO₄ and 0.5N NaOH.

20µg/L (ppb) was the recommended target specified in the extractables testing recommendations by BioPhorum published 2014, which may not be possible for all elements in all matrixes. Data above the LOQ will be reported. In the summary tabe of the BEDS template only results ≥20 ppb are reported.

It is optional to include measurements of pH, TOC and NVR. If included, it is generally sufficient to apply to 50% Ethanol and/or water samples as appropriate.

Additional tests may be requested by an end-user where they deem it necessary for their product/process. In such cases, it will be for the user and supplier to reach an agreement on this testing and the related funding.

Exceptions may arise from time to time. In these cases, an agreement should be reached between the single-use system (SUS) supplier and the customer regarding costs, e.g. the request can be performed as a fee for service by the supplier.

An extractable profile for pure organic solvent is sometimes overkill and not as representative of what is relevant to the processes, which BioPhorum members have been concerned with. Therefore, using a pure organic solvent is not relevant and not needed for our applications.

The BioPhorum protocol published 2014 recommends three extractions with dichloromethane (DCM) at 1:1 (volume/volume) ratio. It is allowed to alter volume ratios or the number of extraction cycles to reach sufficient spike recovery and LOD of the standard(s) used in method qualification.

The volume of the combined DCM fractions will depend on the volume of extract used for LLE Thus, the recommendation from 2014 of evaporating down to ~1mL may not be applicable, depending on the starting volume. It is crucial not to evaporate to dryness, as compounds may be lost or not dissolve again. There is no need to reconstitute the concentrated extract to the original sample aliquot volume.

The selection of the extraction solvents was intended to encompass the widest range of potential extractables relevant to our processes (following the 80/20 approach). The extractable profiles of the solvents are selected to overlap to ensure this coverage is thorough. Therefore, we acknowledge that there are some compounds whose extraction efficiency is improved by pH adjustments to the extraction solvents. However, since the goal is to gain a complete picture of the extractables via all of the solvents and time points, the loss of efficiency in one solvent condition is compensated with a more complete picture by the other solvents. If a laboratory prefers to pH adjust, then that is fine and should be specified within the report. If there are specific extractables that a company expects to be present and are known to require pH adjustment, then we support this approach as long as the adjustment and rationale are detailed in the report.

Refrigeration may be good for the short-term storage of samples. Sample containers should be inert glass, Teflon or stainless steel.

We also recognize that the sample storage issue is important, with many opinions. It is always a point of consideration because the ideal would be real-time analysis but that is not practical. We can see the scientific justification for both freezing and storage at 2–8°C. Precipitation and phase separation are risks when freezing extracts, but storage at 2–8°C may lead to the degradation of extractables (oxidation/hydrolysis, etc.) or loss of volatiles.

Our baseline is that, barring a perfect solution, we leave it to the analytical laboratories to justify their approach as being reasonable.

For a component, a family is a range of components manufactured from the same materials using the same manufacturing process at the same manufacturer but may be of different sizes/shapes. An example is polypropylene connectors molded into different shapes (e.g. L, T or Y) from the same resin. Another example is silicone tubing where the family may be platinum-cured silicone tubing of varying internal diameter measurements. Platinum-cured silicone tubing manufactured by a different manufacturer is not part of the family nor is peroxide-cured silicone tubing. The experimentally obtained extractables data can be applied to other components in the family by a conversion factor based on surface area (or weight).

For filters, a component family can be defined as filters of different sizes manufactured from the same filter membrane type and raw materials in the housing and connectors. The experimentally obtained extractables data can be applied to other components in the family by a conversion factor based on the surface area. This is a simplified conversion model, as extractables originating from the housing and connectors are not related to the membrane surface area and thereby scale differently.

For sensors and valves, a component family is a range of components manufactured from the same materials of construction but in different proportions or different sizes/shapes. The experimentally obtained extractables data can be applied to other components in the family by a conversion factor based on the total surface area of fluid-contacting parts or weight, as appropriate. This is a simplified conversion model, as extractables originating from the different sub-components scale differently.

For bags, the preferred study design is to test each component of the bag assembly individually (e.g. film, ports, tubing, impeller, sensors, filters, etc.).

Similar to a component family, an assembly family can be defined. A bag assembly family is a set of bags manufactured from the same components but in different ratios. Examples are bags that differ in size, number of ports or length of attached tubing. The extractables profile for the assembled bag will be obtained by combining the extractables data for each component.

Alternatively, if an assembled bag was tested as a representative of an assembly family, the experimentally obtained extractables data can be applied to other bag assemblies in the product family by a conversion factor based on the film surface area. This is a simplified conversion model, as extractables originating from the other components are not related to the film surface area.

For chromatography columns, the preferred study design is coupons of each material. A column-assembly family is a set of columns manufactured from the same materials of construction but in different sizes/ratios.

Alternatively, if a column was tested as an assembly, the experimentally obtained extractables data can be applied to other columns in the assembly family by a conversion factor based on the surface area of the main fluid-contacting part(s). This is a simplified conversion model, as the ratio between materials varies with column size.

Bags are assemblies that, besides the film, also include components such as ports, tubing, connectors, filters, sensors and impellers, which all are listed as separate test items in the BioPhorum protocolThus, it is preferred to test the film and the other bag components separately. Impellers may be difficult to scale down and it is thus recommended to test the individual impeller materials separately in the format of coupons.

Vent filters are not considered to be in fluid contact and do not need to be tested.

Time points for film used in storage bags are T1, T21 and T70 days.

Time points for film used in bioreactor and mixing bags are T1 and T21 days.

For non-homogenous/non-symmetric tubing, only the inner surface should be exposed to the extraction solvent. The tubing thus has to be filled with an extraction solvent. For homogenous tubing, it is allowed to fully immerse the tubing in the extraction solvent.

Express the result as µg/cm².

Time points for tubing used on storage bags are T1, T21 and T70 days.

Time points for tubing not intended to be used on storage bags (e.g. tubing in pumps or fluid-transfer kits) are T1 and T21 days.

The general consensus is that the amount of surface area truly impacted by the weld is very small so as to be negligible to E/L profile. The welding process occur typically with the fluid squeezed away from the cutting and welding bend area. Once it is cooled to room temperature, the surface area of the welded bend is very small in comparison to the overall surface area of the single-use system.

For non-homogenous/non-symmetric connectors, only the inner surface should be exposed to the extraction solvent. The connectors thus have to be filled with extraction solvent. Multiple connectors can be used and the extracts pooled before analysis.

Express the results as µg/cm².

The same sensor may be used with a wide range of tubing.

Thus, it is preferred to design the study to only reflect the extractables of the sensor.

For flow-through sensors, it will be necessary to use tubing to achieve enough volume for analysis as only the inner surface of the sensor is wetted. This requires a separate study setup of tubing to subtract the extractables associated with the tubing.

Express the results as µg/cm². When there are no other sensors in the component family the results can be expressed as µg/sensor.

Chromatography columns are, in many cases, large and cannot be scaled down. Thus, it is recommended to use coupons of each individual fluid-contacting material.

Alternatively, entire columns can be filled with an extraction solvent.

Only depth filters using polymeric media are in scope and can be tested according to the category ‘Sterilizing-grade filters/process filters’ as depth filters are commonly used downstream and not only for clarification after cell culturing. Filters using non-polymeric media, e.g. diatomaceous earth, are not in scope.

These processes were out of scope when the BioPhorum protocol was developed. Guidance for these and other processes will be given in the future.

The reason for setting ≥1:1 as the target effective filtration area to volume ratio (EFA/V) for filters is to allow for a higher volume of extraction solvent for recirculation.

We acknowledge that the target surface area to volume ratio (SA/V) of ≥6:1 is difficult to achieve for most test items. Thus, it is accepted to use a lower SA/V than the target, as long as the SA/V arrived at is greater than the SA/V in the component’s intended use.

Yes.

Preferred is to include two separate extractions from components in the same family manufactured from two different resin lots.

If this is not feasible, two separate extractions from two components in the same family manufactured from the same resin lot in two different manufacturing events can be used.

If none of the above options are feasible, two extractions from one lot of component can be used.

The requirement is that the gamma dose should fall within the upper 10 kGy of the dose range specification. However, many sterilization facilities have a regular dose window of 15 kGy. An acceptable option is then to allow the outcome to fall within 10 kGy from the upper specification limit, but also allow it to exceed the upper limit by ≤5 kGy (e.g. 40–55 kGy if the maximum-allowed dose is 40–50 kGy).

An alternative approach is to use a gamma irradiator that has a tighter dose range, such as those typically used for dose audits and validations.

According to the BioPhorum protocol published in 2014 the maximum time between gamma irradiation and initiation of extraction is five weeks, i.e. the extraction should start within five weeks from the gamma irradiation date. This is to avoid a loss of volatile extractables. We acknowledge this is a short timeframe in relation to shipping logistics and have revised the requirement to allow up to eight weeks from gamma date to initiation of extraction.

Inverse staggering of an extraction start is when, for example, the 21 days extraction is started on day 1 and the 24 hours extraction is started on day 20. All samples are then collected at the same time and analyzed on the same occasion.

This approach is allowed as long as the time from gamma to initiation of extraction start does not exceed eight weeks.

It is recommended to gamma irradiate all test items in a single gamma run.

The staggering of gamma is when a lot is split up to be gamma irradiated on multiple dates, e.g. the components for T70 days extraction are gamma irradiated on one occasion and the components for T1 and T21 days extraction are gamma irradiated on a later occasion. This is to be able to collect and analyze all extracts on one occasion, while still fulfilling the requirement of the maximum time from gamma date to extraction initiation. This approach is allowed as long as the gamma dose received on the second occasion is comparable to the dose received on the first occasion.

The assembly product family information template lists components included in an assembly. This template is available on the BioPhorum website.

Preferably, one extractables report should be provided for each component in an assembly. The BioPhorum Extractables Data Summary (BEDS) template is also available on the BioPhorum website.

Alternatively, extractables studies of several components can be merged into one report, if appropriate.

The SUS supplier(integrator is the direct interface with the end-user (i.e. the biopharmaceutical manufacturer) and, as such, is the primary supplier responsible for all aspects of quality for all components relating to the SUS it is selling. This includes assembling the required information regarding the extractables test data in the standardized format identified in the BioPhorum protocol and the Standard User Requirements package sent to each SUS supplier.

When required, the SUS suppliers need to have clear quality agreements in place with their own suppliers, stipulating the requirement for these data. The data can come from the component manufacturer to the SUS manufacturer – the key point is that the data are provided.

Sharing customer names is not a standard business practice, as often biotech component manufacturers can be N-1 suppliers as well as competitors. BioPhorum recommends the ‘data pass-through’ approach whereby the component manufacturer works out an agreement for the primary supplier/integrator to be able to share relevant reports with the end-user purchaser (See BioPhorum website).  This is also consistent with approaches for other existing compendial requirements. In the event the component manufacturer is unwilling/unable to implement the pass-through approach for specific components, then the component manufacturer is expected to inform the integrator/primary supplier what datasets are available and how the primary customer/end-user can easily access.

The bioprocessing industry is governed by cGxP Whether performed internally by the component manufacturer or through an external contract research organization (CRO), any data that are used to make quality decisions must meet minimum data integrity and traceability standards and are expected to be under oversight of a quality management system.  It is the expectation of BioPhorum member companies that data provided by suppliers is generated in such a way as to be compliant with these standards. While it is not necessary for data to be generated to cGxP, it is expected that the component manufacturer (i.e. study sponsor) takes ownership of the study design and review of the data. Typically, this will require technical functional area input and oversight from the supplier’s Quality Unit.

For clarity, where a third party (such as a CRO) has performed extractables testing, it is not sufficient for its report to simply be passed along to the end-user. Information related to the specific component tested and study design should be reported in the standard BioPhorum Extractables Data Summary (BEDS) format and be under control of the component manufacturer’s quality management system.

Yes, the extractables team is forming a community of practice that will meet periodically to review feedback and questions that have arisen through the application of this recommended testing.