Acceptable Intake (AI) for NDSRIs: FDA vs EMA vs Health Canada

When examining the AI limits of different regulatory authorities for nitrosamine, clear differences exist in the regulatory philosophy for setting AI limits. All the major regulatory authorities aim to safeguard patient safety by applying different methods to assess, quantify, and manage the nitrosamine impurity risk. These differences affect the threshold for AI, and as the threshold changes from regulator to regulator, it affects the interpretation of carcinogenicity data. Companies often need separate justifications, testing strategies, and documentation for each region. This increases the workload on pharmaceutical organisations working in multiple regions. The same nitrosamine impurities trigger different compliance requirements across different markets and regulatory authorities.

This article examines the different approaches applied by each regulator and their implications for pharmaceutical manufacturing.

1. FDA:

Overview

FDA considers ICH M7 for the determination of nitrosamine AI limits. The FDA emphasises scientific understanding and structural similarity in the presence of adequate data. FDA encourages manufacturers to propose impurity-specific AI limits rather than relying on fixed thresholds. Rather than enforcing uniformity, FDA accepts transparency, analytical validation, and risk-based decision-making. This AI limit should be supported by sound toxicological reasoning. This approach clearly indicates that the FDA is keen on patient safety while also being aware of practical manufacturing difficulties.

Thus, FDA emphasises flexibility through risk communication. FDA uses a three-tier data system in which:

• Direct rodent carcinogenicity data.
• CPCA and (Q)SAR model prediction.
• Read-across from structurally similar nitrosamines is suggested.

This three-tier system ensures that reliable data is used first. The default AI limit for unknowns is 26.5 ng/day. FDA endorses a five-category system with a range from 26.5 (ng/day) to 1500 (ng/day). FDA may not accept the negative enhanced Ames test and may ask for additional in vitro mammalian cell mutation study data to support the 1500 (ng/day) limit. FDA has published a list of recommended AI limits for more than 260 NDSRIs on its website.

2. EMA:  

Overview

EMA suggests the TD50-based model for the calculation of AI limits. EMA applies a standard but somewhat precautionary approach. EMA aims to reduce variability by applying consistent methodologies across products and geography. EMA suggests that if multiple nitrosamine impurities exist in one molecule, the combined exposure must remain below the lowest AI. EMA requires in-depth process mapping to identify all potential nitrosamine formation pathways. It suggests a default AI limit of 18 ng/day. EMA has published a list of recommended AI limits for more than 130 NDSRIs on its website.

 3. Health Canada:

Overview

Health Canada uses a hybrid approach, combining EMA’s conservative principles with the FDA’s flexibility. Health Canada mostly follows the EMA’s TD50 model, but also allows CPCA or (Q)SAR justification with a sound technical approach and data. This hybrid approach allows a balance between practical implications and patient safety. Health Canada suggests performing lot-to-lot release testing for high-risk products. Health Canada accepts negative Ames test results to justify higher AI limits of 1500 ng/day. It suggests a default AI limit of 18 ng/day, in line with EMA. Health Canada has published a list of recommended AI limits for approximately 130 NDSRIs on its website.

Differences Explained in Tabular Form

Agency	Default AI Limit	Methodology	Category / Range	Primary Data Basis
FDA	26.5 ng/day	ICH M7	5 Category, 26.5 ng/day to 1500 ng/day	Structure–activity (CPCA, (Q)SAR, read-across)
EMA	18 ng/day	TD50-based model	5 Category, 18 ng/day to 1500 ng/day	Carcinogenic potency (TD50)
Health Canada	18 ng/day	Hybrid (ICH M7 + TD50)	AI limits based on data source including CPCA	Read-across; CPCA; (Q)SAR

Responsibility of the Quality Control Team (Documentation):

Bridging Regional Requirements

Given the regional differences described above, QC teams operating across multiple markets must build a documentation strategy that satisfies all three regulatory frameworks simultaneously.

The QC team must document a multistep mitigation strategy to ensure compliance with different regulators for nitrosamine by performing:

• Risk Assessment.
• Confirmatory testing.
• Mitigation plan preparation and reporting.

The QC team must evaluate the risk of formation pathways of all possible nitrosamines in the materials used, such as excipients, solvents, catalysts, raw materials, and synthetic processes. The team should also evaluate the possibility of cross-contamination from equipment used during the manufacturing process. It is advisable to evaluate the product during stability and storage for any change or increase in nitrosamine levels. If nitrosamines are confirmed during the evaluation, a root cause analysis for their formation, along with a control strategy, must be documented. The QC team must plan for continuous monitoring of nitrosamines to check trends through analytical testing as part of the stability programme. A scientific and logical justification with data should be presented to support the proposed AI limits.

Case Study:

Agency Divergence: N-Nitroso-Dorzolamide

There are specific examples where the agencies list different AI limits for the same NDSRI based on their individual assessments. N-nitroso-dorzolamide has an AI of 100 ng/day for the FDA, while EMA and Health Canada suggest 1500 ng/day according to their assessments.

Process Risk Assessment: Miglustat

In the case of Miglustat, the organisation conducted an in-depth risk assessment and identified a risk of nitrosamine formation from excipients such as starch and MCC. Theoretical calculations suggested that these nitrosamines would not increase during stability. This theoretical justification was supported by confirmatory testing. The conclusion of confirmatory testing aligned with this, and it was concluded that routine testing is not required for these nitrosamines.

Conclusion:

From the above discussion and comparison, clear differences exist between the FDA, EMA, and Health Canada for nitrosamine AI limits. The FDA uses ICH M7, EMA enforces TD50 standardised conservatism, and Health Canada follows a balanced hybrid model. Understanding these differences is essential for pharmaceutical companies developing multi-regional submission strategies.

FAQ’s:

Reference:

• United State Pharmacopeia. “🇪🇺 EMA/409815/2020 Rev.21 (19-Jul-24).” Nitrosamines Exchange, 22 July 2024, nitrosamines.usp.org/t/ema-409815-2020-rev-21-19-jul-24/10720.
• Canada, Health. “Nitrosamine Impurities in Medications: Guidance – Canada.ca.” Canada.ca, 2022, www.canada.ca/en/health-canada/services/drugs-health-products/compliance-enforcement/information-health-product/drugs/nitrosamine-impurities/medications-guidance.html.