Where large excesses of reagents are concerned (e.g. >2 eq.), one can generally assume 1 equivalent has been used for reactions that produce > 90% yield of product. For reaction where < 90% conversion is observed, the assessors should consider including the residual reagent that may be remaining through incomplete conversion to product. If stoichiometric or only minor excesses of reagent are used, then it is considered best practice to base the equivalents used, and therefore equivalents left behind, on the yield of product. In this case study, yields were generally significantly over 90% and the decision was taken to assume 1 equivalent of reaction for the reagent of concern.
Given the permitted limit based on a 600 μg/day dose is <2500 μg/g within GW641597X, the required purge was conservatively calculated on the basis of this potential 1 mol excess of 2, which leads to a purge factor of 400 being required for 2 (1 000 000/2500 = 400).
Ethyl bromoisobutyrate 2 reacts to form the ester 3 in the Stage 1a process. An excess of this reagent is required because 2 polymerizes during the process. A conservative approach was taken such that no reactivity for this side process was assigned and neither was any reactivity of this reagent anticipated within the oxidative Stage 1b process. No solubility can be predicted during the stages to prepare 3 and 4 as there is no formal product isolation stage.7 Moderate reactivity is predicted within the process to alkylated 4 with 8 as there is opportunity for alkyl bromide 2 to react with the phenolate derived from 4, as well as potential to form ethers from reaction with the ethanol solvent during the manufacture of ester 9. Ethyl bromoisobutyrate 2 is anticipated to react with the aqueous base during the preparation of crude GW641597X, through hydrolysis, polymerization, and potential ether formation. As 2 is an oil, full solubility was anticipated during the isolation and washing processes during the preparation of crude GW641597X and its subsequent purification. Based on this highly conservative approach, the purge ratio (Table 3.7) for 2 is insufficient to justify an ICH M7 Option 4 control strategy by itself as the purge ratio is less than 1000 [19] and, without further information to support an Option 4 control strategy, an alternative ICH M7 control option would be recommended. One option would have been to measure the level of 2 at the completion of Stage 1. Other options can include the use of “prior knowledge,” which can be literature reactivity data for similar transformations, with structurally similar compounds, in order to further justify/reinforce an Option 4 proposal. Such data can also be used to assign a higher purge score (reactivity, solubility, etc.) if the original predictions are considered overly conservative when reviewed against a similar reported transformation.
Table 3.7 Purge predictions for ethyl bromoisobutyrate 2.
| Stage | Reactivity | Solubility | Volatility | Total | Rationale |
|---|---|---|---|---|---|
| 1a | 1 | 1 | 1 | 1 | 2 M equivalents used where 1 is used to prepare ester 3. Telescoped process; therefore, no isolation. It was also decided to assume no additional reactivity; therefore, 1 equivalent remains at the end of this stage (conservative). |
| 1b | 1 | 1 | 1 | 1 | No reactivity or solubility predicted from Stage 1b during the preparation of 4. |
| 4a | 10 | 1 | 1 | 10 | Expected to react with the phenolate from 4 during Stage 4a to prepare ester 9, potential etherification from reaction with ethanol solvent and polymerization. |
| 4b | 100 | 10 | 1 | 1000 | Anticipated to react with the aqueous base through hydrolysis, polymerization, and potential etherification during the hydrolysis of 9 (Stage 4b). Ethyl bromoisobutyrate 2 is an oil and highly soluble in the isolation solvent. |
| 5 | 1 | 10 | 1 | 10 | Ethyl bromoisobutyrate 2 is an oil and likely to be highly soluble in the isolation solvent. |
| Predicted purge factor (Required purge) | 1 × 105 (400) Purge ratio = 250 (would inform ICH M7 Option 1, 2, or 3 control without supporting data, e.g. collection of supporting “non‐trace” experimental data) | ||||
3.6.1.2 Hydroxylamine
Hydroxylamine reacts with benzonitrile 6 during Stage 2 to form the benzimidamide 7 (Figure 3.4). A 2.5 molar excess of the reagent is used to assure reaction completion, and therefore, as much as 1.5 equivalents might remain within the Stage 2 process after formation of 7.
Hydroxylamine has an ADI of 23 μg/day that equates to a permitted limit of ≤ 38 333 μg/g of GW641597X and correlates to a required purge factor of 39.1 (1 500 000/38 333 = 39.1).
With respect downstream purging, hydroxylamine free base has a low boiling point (58°C), and it is reasonable to expect losses through evaporation during higher temperature processing as well as during isolation and drying through volatilization. Despite this, a conservative approach was taken and volatility was not scored for the entire process. Hydroxylamine can be considered to purge through solubility within processes to prepare 7, 8, 9, and GW641597X, and reactivity was anticipated during Stages 3 and 4 for the preparation of 8 and 9 through reaction with chloroacetyl chloride and alkyl chloride 8, respectively. Even with the application of a highly conservative approach i.e. assuming a starting concentration of 1.5 equivalents and discounting likely volatility, the predicted purge factor was significant, 1 × 108 (Table 3.8). The purge ratio is significantly greater than 1000, which justifies an ICH M7 Option 4 control rationale without recourse to additional experimental data and simply reporting “unlikely to persist” should be sufficient for a regulatory submission as per the published guidance [19].
Table 3.8 Purge predictions for hydroxylamine.
| Stage | Reactivity | Solubility | Volatility | Total | Rationale |
|---|---|---|---|---|---|
| 2 | 1 | 10 | 1 | 10 |
|