Barriers to the Use of Recombinant Bacterial Endotoxins Test Methods in Parenteral Drug,Vaccine and Device Safety Testing

Abstract

The Bacterial Endotoxins Test (BET) is a critical safety test that is used to detect bacterial endotoxins, which are the major contributor to fever-inducing contamination risks known as pyrogens. All parenteral therapies, including every lot of injected drugs, vaccines, medical devices, must be tested for pyrogens to ensure patient safety. Bacterial endotoxins test methods were developed as a highly sensitive detection method for bacterial endotoxins, after the discovery of a clotting cascade in horseshoe crab blood. However, horseshoe crab species are limited to some inshore coastal habitats along the Atlantic coast of the USA and others throughout Asia. Fully functional horseshoe crab clotting factors can be manufactured via recombinant protein production, and several BET methods featuring recombinant horseshoe crab proteins have now been developed for commercial use. Recombinant Bacterial Endotoxins Test (rBET) methods based on the use of recombinant Factor C (rFC) were established in the European Pharmacopoeia — however, these methods have not yet been granted compendial status in the United States Pharmacopoeia (USP). In order to facilitate dialogue between stakeholders, the Physicians Committee for Responsible Medicine hosted two virtual roundtable discussions on the perceived barriers to the use of rBET methods for US FDA requirements. Stakeholders agreed that multiple rFC-based methods have been demonstrated to have suitable analytical performance, as described in ICH Q2 on the Validation of Analytical Procedures and USP <1225> on the Validation of Compendial Procedures. United States Pharmacopoeia compendial inclusion of the rFC-based and other rBET methods was favoured, in order to reduce the additional burdens created by a lack of global harmonisation on BET testing requirements.

Executive summary

The Bacterial Endotoxins Test (BET) is a critical safety test used to detect bacterial endotoxins, the major contributor to fever-inducing contamination risks known as pyrogens. All parenteral therapies, including every manufacturing lot of injected drugs and vaccines, as well as medical devices, are tested for pyrogens to ensure patient safety. BET methods were developed after the discovery of a clotting cascade in horseshoe crab (Limulus polyphemus and Tachypleus spp.) blood as a highly sensitive detection method for bacterial endotoxins. Notably, the horseshoe crab is not a species of crab, but rather an extraordinarily unique arthropod that is limited to some inshore coastal habitats along the Atlantic coast of the USA and others throughout Asia. Recent surveys by the International Union for Conservation of Nature (IUCN) have found the major horseshoe crab species in Asia and the USA to be Endangered and Vulnerable to extinction, respectively.

Fortunately, fully functional horseshoe crab clotting factors can be manufactured via recombinant protein production, and several BET methods featuring recombinant horseshoe crab proteins have now been developed for commercial use. Recombinant BET methods based on the use of recombinant Factor C (rFC) have been established in the European Pharmacopoeia (Ph. Eur.) since January 2021. However, these methods have not yet been granted compendial status in the United States Pharmacopoeia (USP), and are still listed as “alternative methods”. In order to facilitate dialogue between stakeholders on this issue, the Physicians Committee for Responsible Medicine hosted two virtual roundtable discussions on the perceived barriers to the use of recombinant Bacterial Endotoxins Test (rBET) methods for US Food and Drug Administration (FDA) requirements to ensure the safety of medicines and medical devices. This workshop report contains summaries of both roundtable discussions. The first, focusing on rBET methods based on the use of rFC, was held virtually on 3 March 2021; the second, which covered rFC and the newly-developed recombinant Cascade Reagent (rCR) method, was held virtually on 28 June 2021. The first roundtable included 22 participants from pharmaceutical companies, rFC suppliers, regulatory agencies, NGOs, a trade group, the USP, and consultants. The second roundtable included 37 participants from pharmaceutical companies, rFC suppliers, Limulus Amoebocyte Lysate (LAL) suppliers, regulatory agencies, NGOs, a trade group, the USP, and consultants. Further details are included in the Acknowledgements section.

While roundtable participants representing different commercial interests could not come to full agreement on specifics, some areas of common ground were identified for consideration in moving toward global harmonisation of regulatory BET requirements. Participants widely agreed that rBET methods represent an ideal solution to the need to protect horseshoe crabs, and also that their use cannot be implemented if there is a threat to patient safety. Stakeholders agreed that rFC-based methods have been demonstrated to have suitable analytical performance, as described in ICH Q2 on the Validation of Analytical Procedures and USP <1225> on the Validation of Compendial Procedures, while the newer rCR methods have shown promise in preliminary evaluations. While it was noted that companies can submit rBET results for BET requirements under current FDA guidance — and multiple therapies have already been approved by the FDA for market release using rFC-based results through this mechanism — a lack of global harmonisation on BET testing requirements creates an additional resource burden. USP compendial inclusion of rBET methods was favoured, in order to reduce the additional validation steps required for non-compendial BET methods. Compendial status also reduces uncertainty in regulatory submissions, by providing clear guidance for acceptance.

Notably, since these roundtable discussions were convened, the USP has dismissed its previous Microbiology Expert Committee and asserted a commitment to making progress on endotoxins testing with a new expert committee in 2023. In August 2023, after evaluating the data and science, the new expert committee proposed a new standard for evaluating BET using non-animal derived reagents, which would include rFC-based methods and rCR-based methods. ¹ The USP welcomes comments on this proposal by 31 January 2024. The authors therefore further recommend that the new expert committee act with urgency to review stakeholder input and expedite the inclusion of rBET methods as compendial methods within the USP.

Background

All vaccines, injectable therapies and medical devices, as well as all raw materials and packaging involved in their manufacturing and administration, must be tested for fever-inducing contaminants (i.e. pyrogens) before they can safely be administered to humans. The Bacterial Endotoxins Test (BET) method was developed after the discovery of a clotting cascade in horseshoe crab (Limulus polyphemus and Tachypleus spp.) blood, now known as the Limulus Amoebocyte Lysate (LAL) and Tachypleus Amoebocyte Lysate (TAL) assays. ² BET methods provide rapid and sensitive analysis of the most common pyrogen risks during development and production. As a result of these advantages, the BET has largely replaced the Rabbit Pyrogen Test following its regulatory acceptance by the FDA via USP compendial inclusion in the 1980s. LAL-based BET methods have since been the default tests used by industry for pyrogen testing in the USA. However, BET methods featuring recombinantly produced horseshoe crab clotting protein(s) are now commercially available from multiple suppliers. This is fortuitous, as alternative BET reagents are urgently needed to end the commercial exploitation of horseshoe crabs without jeopardising patient safety.

With continuous increases in global production of critical parenteral medicines, such as vaccines, injectable therapies, and medical devices, the demand for BET testing has been a major contributor to the exploitation of this ‘living fossil’ which has occupied its ecological niche relatively unchanged for over 400 million years. ³ Globally, there are four extant species of the horseshoe crab — one along the Atlantic coast of the USA (Limulus polyphemus) and three throughout Asia (Tachypleus tridentatus, T. gigas and Carcinoscorpius rotundicauda). T. tridentatus, the major Asian species, was recently listed by IUCN as Endangered after a recent assessment, as a result of habitat loss and overfishing. ⁴

The solitary American species, representing the only major horseshoe crab population remaining worldwide, was last assessed by the IUCN in 2016 and listed as Vulnerable to extinction at that time. ³ In the USA, biomedical harvest of Limulus for LAL production has steadily increased year on year, since 2016. ⁵ While estimates vary, the accompanying direct mortality of horseshoe crabs is estimated to be 15%. ⁵ Survivors suffer from behavioural and physiological deficits that could negatively affect future generations, including reduced participation in spawning activity and sustained reduction in haemocyanin.^6,7

Fortunately, the genes encoding the proteins in the horseshoe crab clotting cascade have been sequenced, and fully functional cascade proteases can be produced via recombinant protein production. Recombinant protein expression involves the insertion of the gene of interest into the genome of host cells to produce large quantities of the protein of interest for subsequent isolation from cell lysates. Recombinant protein production is behind numerous biomedical innovations of the late 20th century, and has various benefits — including eliminating disease risks associated with therapeutic peptides and proteins derived from animals. The gene encoding Factor C of the horseshoe crab clotting cascade was the first to be cloned and sequenced for the production of fully functional recombinant protein. ⁸ Subsequently, the genes encoding the other cascade proteases Factor B, pro-clotting factor and Factor G were also identified, sequenced and successfully expressed in host cells.^9,10

While Factor C is sufficient to detect endotoxin, rFC-based methods require a fluorescent detection method to maintain sensitivity without the amplification of the full cascade. In 2021, a new recombinant BET method featuring the three endotoxin-specific cascade enzymes (excluding the glucan-sensitive Factor G) was introduced and marketed as recombinant Cascade Reagent (rCR). This method maintains sensitivity with the same detection method as chromogenic LAL-based methods. Currently, several BET methods featuring recombinant horseshoe crab protein(s) are available for commercial use. However, their adoption for the safety testing of drugs and devices is regulated indirectly through published pharmacopoeias, including the US Pharmacopoeia (USP), the European Pharmacopoeia (Ph. Eur.), as well as pharmacopoeias serving Japan, China, Korea, Brazil, Argentina, and many other countries.

To help maintain the quality of medical products, pharmacopeias include chapters describing which test methods may be used in product development and manufacturing for different safety concerns, such as endotoxin contamination. A method included as compendial in a pharmacopeia means that a product developer can use the method with full confidence that the results will be accepted by the relevant local medical regulatory agency. Those that wish to use a method not included as compendial may work with regulatory agencies to make a case-by-case determination of acceptability by providing additional evaluation studies each time they seek to use the method. The International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH) details specific expectations for performance validation of analytical procedures; quantitative assays must demonstrate specificity, suitability of calibration within the working range, accuracy and reproducibility/precision. Repeating this validation for every manufactured product is time- and cost-prohibitive.

The Ph. Eur. now includes both LAL-based and rFC-based methods as BET methods, with the addition of Chapter 2.6.32, Test for bacterial endotoxins using recombinant Factor C. ¹¹ This is part of a wider effort within the European Directorate for the Quality of Medicines (EDQM) to establish compendial non-animal test methods for both endotoxins and non-endotoxin pyrogens. ¹² The USP considered a modification to USP <85> on the Bacterial Endotoxins Test, in 2019 ¹³ to include recombinant reagents. However, this proposal was rescinded in 2020 and USP <1085.1> on the Use of Recombinant Reagents in the Bacterial Endotoxins Test was instead drafted and made available for public comment. ¹⁴ Numerous stakeholders criticised the appropriateness of the proposed requirements in this proposed guidance chapter. ¹⁵ The guidance chapter does not equate to equivalent compendial inclusion of rFC reagents for BET, and has not yet been published at the time of this report. To the authors’ knowledge, the newly developed rCR method has not yet been assessed for compendial inclusion.

Recombinant Factor C (rFC) Roundtable Discussion

Participants’ interests in engaging on the compendial status of rFC-based BET methods

Initial roundtable discussions revealed that several pharmaceutical companies are currently using rFC-based BET methods, although only one company indicated that it was currently using an rFC-based BET method to ensure the safety of US-marketed products. All other companies indicated that they wanted to participate in the discussion because they were at various stages of evaluating the feasibility of adopting rFC-based methods for their BET needs. Many participants stated that their interest in using rFC-based BET methods was due to the multiple benefits of these methods, when compared to LAL-based BET methods. A number of participants reported that switching from LAL-based to rFC-based BET methods aligns with their company’s ethical commitments to replace animal use in accordance with the Three Rs principles (replacement, reduction and refinement of animal use), and/or as part of corporate sustainability goals. Other participants noted that the supply of recombinant protein, as a scalable synthetic option, avoids supply chain vulnerabilities associated with dependence on animal blood. For some companies, adopting an rFC-based BET method was described as a proactive buffer against any reasonably foreseeable price hikes resulting from the limited sources of LAL. Scientific advantages were also noted, including that rFC-based methods use a defined enzymatic reagent instead of a variable lysate, and that the rFC assay principle avoids the falsely elevated readouts that can arise from non-specific pathway activation in LAL assays.

Scientific considerations for the adoption of rFC-based BET methods

The meeting organisers opened the discussion by asking participants to share any concerns about the technical performance of rFC-based BET methods. The primary performance concern of rFC-based methods was identified as the conflicting interpretations of comparability studies. It has been asserted that analytical comparisons of characterised endotoxin standards are insufficient to establish comparable performance between LAL-based and rFC-based BET methods, and that comparable performance must also be shown for indigenous (‘autochthonous’) endotoxin samples that arise from contaminated manufacturing processes, before rFC-based BET methods can be considered equivalent to LAL-based methods.^16–18

In this published opinion, co-authored by several members of the 2020 USP Microbiology Expert Committee, it was argued that due to the wide variety of endotoxins that might be encountered in manufacturing processes, and how they might differ from the analytical standards used to calibrate BET assays, additional comparisons of ‘autochthonous’ endotoxin samples are necessary to ensure safety.^16–18 Several participants in the roundtable discussion noted published studies that consistently show comparability of rFC-based and LAL-based BET methods for quantifying endotoxin in ‘autochthonous’ samples, and/or demonstrate advantages to the use of rFC-based methods.^19–25 In addition, it was noted that a recent study concluding that rFC is not ready for use ²⁶ was published by an LAL supplier. Multiple participants expressed concern that LAL suppliers have previously published similar claims in journals known to deviate from best editorial and publication practices.

Several participants stated that they have data from their own internal comparability studies, both published and unpublished, with some of these participants indicating that they are currently working with the USP to provide data from these comparison studies. Industry participants expressed interest in working collaboratively with the USP Microbiology Expert Committee to provide additional data. However, it was felt that the specific data requirements that are needed for compendial status approval should be made explicit. Participants also emphasised the need to include both the intra-assay and inter-assay variability that is inherent in the various compendial LAL assays, in all comparisons. This was noted as important, so that any observed differences in results between the rFC-based and LAL-based methods can be interpreted within the context of inherent BET assay variability. In other words, it is necessary to evaluate how quantitative differences between the rFC-based and LAL-based methods compare to quantitative differences between the various compendial LAL-based methods.

Participants also stressed the need for attention to glucan and other assay-interfering complex carbohydrate content in uncharacterised samples, that could nullify conclusions drawn from the study design. LAL-based BET methods are not specific to endotoxin, and elevated signals and false positives can be caused by the presence of certain complex carbohydrates due to other factors present in amoebocyte lysate. ²⁷ It was noted that uncharacterised ‘autochthonous’ manufacturing samples containing cellulosic materials would give rise to elevated readings in LAL-based methods, so it cannot be assumed that LAL-based BET methods would provide an accurate response for these samples.

Finally, the process of pharmacopeial adoption of LAL as an in vitro replacement of the Rabbit Pyrogen Test (RPT), and the introduction of USP chapter <85> to establish the BET, was proposed as a model of expectations for the evaluation of rFC-based BET methods. However, it was noted by many that the situations were not analogous, because recombinant BET methods represent a reagent/protocol modification to the established BET method without introducing a novel assay principle. It was also noted that the adoption of multiple modified reagents and protocols from the original coagulation-based LAL assay to the kinetic chromogenic LAL assay did not require decades of comparability data in order to update the established USP chapter <85> to reflect technological advances in the analytical methods used for BET.

Policy considerations for the adoption of rFC-based BET methods

As with the scientific discussion, the meeting organisers opened this discussion by asking participants to share how current policy affects their ability to implement rFC-based BET methods. Much of the policy discussion centred on the need for harmonisation and USP compendial inclusion of the rFC-based methods, although participants did not agree whether these methods should be included in USP <85> or in a separate chapter. Representatives from companies, especially those further along in the process of in-house validation, expressed an interest in using rFC-based methods for regulatory submissions, but stated that they are limited by the lack of inclusion of the rFC-based methods as compendial BET methods in the USP. Some participants reported that the lack of global harmonisation prevented their company from using rFC-based BET methods; representatives of global companies expressed a need for standards that are consistent across markets. While rFC-based methods are compendial in the Ph. Eur., currently, LAL-based methods are the only BET methods compendial across global markets. This inconsistency creates an extra resource burden when employing rFC-based BET methods that does not exist when using LAL-based BET methods. In light of the focus on COVID-19 throughout 2020 and 2021, multiple companies indicated that time was the major limiting resource in adopting new testing and manufacturing processes.

At least one company has laid the groundwork for rFC-based method acceptance, because the benefits were considered worth the additional resource investment under the current FDA Guidance for Industry. ²⁸ This guidance provides an outline for the additional studies needed for the FDA to approve non-compendial pyrogen and endotoxin tests — indeed, the FDA has approved multiple therapies for market release tested solely with an rFC-based method for BET requirements, where companies have worked with the agency to perform this additional validation. The FDA’s Manual of Policies and Procedures on Acceptability of Standards for Alternative Compendia was suggested as a potentially feasible mechanism for submitting rFC-based assay data with regulatory filings under the current USP. This guidance could allow companies to use an rFC-based method for BET requirements under Ph. Eur. Chapter 2.6.32 in FDA submissions, but it was noted that following this guidance could represent an additional burden in time and resources relative to submitting data from a compendial assay for both sponsors and FDA reviewers. Without USP compendial status, or FDA guidance that clearly communicates that an rFC-based method will be accepted, companies are left to take on the calculated risk of submitting data from a non-compendial BET method without prior assurance that it will be accepted.

Participants then led the discussion to focus on concerns associated with the USP’s guidance chapter proposal for rFC. Technical issues with the draft for USP <1085.1> were voiced by multiple participants, especially with regard to the inclusion of USP <1223> on the Validation of Alternative Microbiological Methods, as necessary for the validation of recombinant BET methods. It was noted that USP <1223> provides guidance on the validation of methods which seek to identify and/or quantify the presence of live microorganisms, which would neither include LAL nor recombinant BET methods. The lack of relevance of pre-filtered water analyses for final product safety was highlighted by a number of participants, as it was indicated that analysis of these samples is not sufficient to identify endotoxins that can be introduced via excipients or other sources. As a response to feedback on the originally proposed changes to USP <85>, a majority of participants favoured a standalone chapter for rFC-based assays, analogous to the recently adopted Ph. Eur. Chapter 2.6.32 to promote global harmonisation efforts. It was noted that the proposed USP <1085.1> draft in PF 46(5) is an informational chapter, and would therefore not fulfil this need. Participants enquired about the next step forward and learned that there was no formal timeline in place — but if the proposed USP <1085.1> were published, it would be accompanied by a response to public comments.

Considerations for the US Pharmacopoeia

Much of the scientific discussion reflected individual conclusions that rFC-based BET methods are, at least scientifically, equivalent to LAL-based BET methods, and existing data demonstrate that rFC should be included as a compendial method in the USP. It was noted that a compendial chapter separate from the current BET chapter (USP <85>) could bring USP back into harmonisation with the Ph. Eur. As a reflection of the scientific discussions, multiple industry participants expressed that, for their company to be able to adopt the rFC-based methods, policymakers need to rely on relevant published data and proceed with the addition of rFC-based methods as compendial BET methods. It was suggested that the USP may benefit from hearing from the Ph. Eur. commission on how it evaluated comparability of rFC-based and LAL-based BET methods. The Harmonization Pharmacopeia Discussion Group was proposed as a possible forum.

Recombinant Bacterial Endotoxins Test (rBET) Roundtable Discussion

The highlights from the recombinant Factor C (rFC) Roundtable Discussion were summarised for all attendees, before continuing with a discussion focusing on comparability, existing data and remaining gaps. The scope of this further discussion included all LAL-based BET assays described in USP <85>, from the initial ‘gel-clot’ LAL assay to turbidimetric and kinetic chromogenic LAL assays, and focused on commercially available rFC and the newly-introduced recombinant Cascade Reagent (rCR) method. It was noted that stakeholders agree that recombinant BET assays have demonstrated robust quantitative validation, as described in the ICH Q2 and USP <1225> guidance.^17,29 However, there was still disagreement among stakeholders as to whether the validation to date is sufficient for establishing the ‘comparability’ of recombinant assays relative to their LAL-based predecessors. These disagreements include how to address the variability of BET assays, endotoxin sources, and the samples to be analysed for comparison.

Process for establishing BET method comparability

The meeting organisers opened a discussion to identify the relevant guidelines for establishing the comparability of BET assays. One participant noted that the term ‘comparability’ in parenteral manufacturing is typically used to refer to products before and after a manufacturing process change, and is not typically applied between methods. Another participant mentioned that comparability is described in the pharmacopeial guidance chapter on the Validation of Alternative Microbiological Methods, USP <1223>. There was marked disagreement among stakeholders on the relevance of this USP chapter for evaluating BET method comparability. Two participants described USP <1223> as the relevant guidance chapter, referring to USP <1223> as a general information chapter that indicates comparability with regard to equivalency. Conversely, a number of participants disagreed with this assertion, claiming that USP <1223> is not relevant because it addresses alternative tests to identify live microorganisms. These participants voiced concern that neither LAL-based nor rBET-based methods can be used to detect live microorganisms, and stated that USP <1225> on the Validation of Compendial Procedures is in fact the relevant guidance chapter on the validation of non-compendial BET methods.

Precedents for establishing comparability of BET assays were then discussed. Participants agreed that exhaustive comparability was shown between the Rabbit Pyrogen Test and LAL ‘gel-clot’ method, resulting in the introduction of USP <85> to describe this novel method. However, the extensive comparability analysis that was performed for the original qualitative ‘gel-clot’ LAL assay was not performed for quantitative LAL assays that were later added to the compendial BET chapter, i.e. USP <85>. It was suggested that the turbidimetric assay was presumed to be equivalent to the ‘gel-clot’ method, as it was added to USP <85> without additional extensive comparability studies. One participant noted that the justification may have been that it is a quantitative measurement of the original coagulation-based assay without synthetic reagent additions. However, other participants noted that the chromogenic method, featuring the introduction of a synthetic colorimetric substrate, was also added to USP <85> without extensive comparability studies. In light of this, many participants expressed concern that recombinant BET methods are being held to a higher standard than LAL-based methods, as more comparability studies are being requested to become compendial, despite operating on the same fundamental principle of endotoxin detection as the original ‘gel-clot’ method.

Existing comparability data

The meeting organisers then opened a discussion to address what has been established with regard to the comparability of rBET assays to existing LAL assays. Many participants from the first roundtable reiterated their views that the previously referenced studies are sufficient for compendial inclusion of rBET methods. Three comparability studies were discussed at length among participants: Firstly, a 2013 study was described as being ideal to demonstrate comparability between LAL and recombinant assays, with general agreement among participants. ¹⁹ This study showed comparability for about a dozen samples, except where samples contained interfering substances that led to the non-specific activation of LAL, which was demonstrated with control assays. One participant wanted to see more studies modelled after this study.

A pair of comparability studies, including endotoxins from numerous bacterial species as well as ‘Naturally Occurring Endotoxins’ (NOEs) and pre-filtered water samples,^20,30 were suggested as additional model comparability studies demonstrating the equivalence of rFC-based and LAL-based BET methods. However, participants did not agree on how to interpret the results. Two participants posited that NOEs is a misnomer when referring to laboratory-derived samples, and that ‘Natural Environmental Endotoxins’ (NEEs), which have changed to adapt to their environment, were the ideal samples for analysis. Multiple stakeholders disagreed, on the basis that environmental adaptation cannot be reproduced across different locations and therefore cannot be used as a reliable industry standard to compare the performance of BET assays. Furthermore, two participants argued that, in addition to NOEs, NEEs were already well-represented in comparative studies. Others contended that neither NOE nor NEE samples are required to validate any other methods approved by the USP Microbiology Expert Committee, such as the Rapid Sterility Test.

A recent study published by an LAL supplier ²⁶ was particularly controversial among participants. Several participants questioned the lack of consideration of the variability of LAL-based BET methods in the study design. There was also significant disagreement among participants about the relevance of the samples evaluated, and participants noted that the editor of the European Journal of Pharmaceutical Sciences issued a ‘statement of concern’ about the conclusions drawn in this paper based on the samples chosen. ³¹

Considerations for future comparability studies

The meeting organisers then opened a discussion for participants to propose what additional studies could contribute to the literature with regard to the comparability of rBET methods. Several participants noted that endotoxin-positive samples are rare. For example, it took several years to accumulate approximately a dozen samples with detectable levels of endotoxin in the study by Chen and Mozier. ¹⁹ A number of participants agreed that studies with large sample sizes, where many samples are tested by multiple laboratories in a range of LAL and recombinant assays, were not practically feasible due to limited relevant samples with detectable endotoxin. Participants also acknowledged that the scale of resources needed to perform multiple tests with the necessary dilutions and controls would be daunting, while providing limited additional benefit. Participants were unable to come to an agreement about what samples, in addition to endotoxin standards, would constitute acceptable surrogates for endotoxin-positive samples, without introducing bias through non-specific activation of Factor G in LAL. One participant reiterated that in the study by Chen and Mozier, LAL-to-LAL comparisons were found to have greater variability than LAL-to-rFC comparisons. ¹⁹ Another participant posited that differences in the preparation of amoebocyte lysates and recombinant reagents could be the rationale for additional comparability studies to evaluate rBET assays. Others argued that rFC-based methods, being less variable than LAL-based methods and more endotoxin-specific, should be considered the standard.

Participants widely agreed that patient safety is of utmost importance. Some participants expressed concern that the intra-assay variability among BET assays were not considered to be a safety risk when evaluating different LAL assays, as this is mitigated by side-by-side analyses when switching between LAL-based methods. One participant reminded attendees that the purpose of the BET is to identify endotoxin hazards, and that minor amounts of quantitative variability between BET assays has not been practically consequential in terms of patient safety, even when statistically significant. Participants agreed that risks from assay variability should not be considered to be specific to recombinant BET assays and should apply equally to LAL and rBET assays.

Conclusions and policy recommendations

In the first roundtable discussion, participants concluded the meeting with clear statements that USP compendial inclusion of rFC-based BET methods is warranted because these methods are at minimum scientifically equivalent to LAL-based BET methods. Multiple advantages of rFC-based methods were identified, namely that they:

— helped to meet organisational commitments to reduce animal use;

— were amenable to rapid scale-up, to avoid supply vulnerabilities;

— represented a protective buffer against reasonably foreseeable cost increases while improving reproducibility; and

— avoided the falsely elevated readouts associated with LAL-based BET methods.

The USP Microbiological Expert Committee’s request for comparability data with indigenous ‘autochthonous’ samples was discussed at length. It was generally agreed that, if a concerted industry effort is to supply the USP with additional data, in order to get to the point of acceptance as a compendial method, then the USP must provide explicit details about how many and what types of samples must be analysed to merit acceptance.

In the second roundtable discussion, participants did not agree on the sufficiency of existing comparability data, nor was a full agreement established regarding best practices for future comparability study design. In the absence of a definitive guideline for BET method comparability, stakeholders were interested in establishing defined expectations on the number of samples that would demonstrate adequate comparability of rBET methods. Some participants also expressed interest in a future discussion with a similar stakeholder group on the acceptance of recombinant Cascade Reagents in the Ph. Eur., as the new Ph. Eur. chapter is only currently applicable to rFC-based BET methods.

Ultimately, with competing commercial agreements at the table, participants did not come to agreement on specifics for objective evaluation of rBET methods for USP inclusion as a compendial BET method. However, several important areas of agreement that were highlighted during these discussions, can be taken into account by stakeholders moving forward:

1. Both the horseshoe crab and the BET play irreplaceable roles in their ecological and economic networks, respectively.

2. The global dependence on Vulnerable and Endangered species for the safety of parenteral drugs and medical devices is not sustainable.

3. rBET methods represent an ideal solution which cannot be implemented at the expense of patient safety.

4. rFC-based BET methods have already been demonstrated to meet regulatory performance expectations for quantitative analytical methods as described in ICH Q2 on the Validation of Analytical Procedures and USP <1225> on the Validation of Compendial Procedures, and have been approved by global health authorities for use.

5. Global harmonisation of BET requirements reduces regulatory burden for pharmaceutical manufacturers. Compendial status for rFC-based BET methods is a crucial step toward re-establishing global harmonisation for BET.

6. The rCR-based BET method has so far shown similar performance to established BET methods and should be established as compendial in global pharmacopoeias as soon as safely possible.

As emphasised in public comments from industry and NGO stakeholders on USP activities, there is a critical urgency to adopt recombinant methods, as parenteral therapies continue to be developed and produced to combat variants in the SARS-CoV-2 pandemic, and global horseshoe crab populations remain in jeopardy. The Ph. Eur. Commission has taken a commendable stand to advance non-animal replacements for pyrogen and endotoxins testing. The authors hope that engaged stakeholders will continue to encourage an open and transparent dialogue about the USP Microbiology Expert Committee’s deliberations and procedures for updating USP <85> or adding an additional compendial chapter on rBET methods.

The authors note that since these roundtable discussions were convened, the USP has dismissed the 2020–2025 Microbiology Expert Committee. ³² In this announcement, the USP espoused a commitment to making progress on endotoxins testing with a new expert committee to complete the five-year term. In August, 2023, after reviewing the data and science, the USP’s new Microbiology Expert Committee proposed a new standard, chapter <86>, that would include rFC and rCR-based methods. The USP invites comments until 21 January 2024.