Chemistry for Life & Health

Controlling Impurities in Drugs

by Will Watson

September 19, 2016

One of the most important aspects of designing, developing, and optimizing the synthesis of an active pharmaceutical ingredient (API) is the control of impurities formed during the reactions. The key is identifying, as far as possible, the structures of the impurities and then working out how they might have formed.

The simplest option is to adjust the reaction conditions (reagents, solvents, temperature, concentration, order of addition, etc.) to minimize or even eliminate impurity formation. Some impurities can never be completely prevented or removed, so the question then becomes, what is their fate? More importantly, can impurities be purged during the downstream steps?

Sometimes an impurity may not be removed immediately, but it may be carried through one or more steps to a point where it can be efficiently rejected. For critical impurities such as mutagenic impurities (MIs, also known as genotoxic impurities, GTIs), strict limits may be placed on how much of the impurity is tolerated in a starting material or key intermediate.

Two recent papers describe developing routes to dolutegravir, an integrase strand transfer inhibitor for treating HIV infections, and vismodegib, a hedgehog pathway inhibitor for treating basal cell carcinoma. In both, the authors pay particular attention to impurity control strategies.


In the first article, Pramod Kumar and co-workers at Micro Labs (Bangalore, India) describe synthesizing the sodium salt of dolutegravir (1 in Figure 1). 

Figure 1

In the first step, the conversion of 2 to 3, two impurities (6 and 7) were formed; they are shown in Figure 2. Optimizing the reaction conditions minimized the formation of 2 to 3, but the impurities were most effectively purged after amide formation in the second step by crystallizing 4. Crystallization also eliminated impurity 8, which arose during step 2.

Figure 2

Three more impurities (911) formed during the synthesis of intermediate 5; they are also shown in Figure 2. As above, 9 and 10 could be minimized by carefully optimizing the reaction conditions, but impurity 11 was much more difficult to remove.

All attempts to control the formation of 11 or to remove it by purifying compound 5 were unsuccessful. So the authors undertook an unusual strategy: They selectively protected 11, a primary alcohol, to form a more soluble molecule. Standard ester formation (benzoate or acetate) was successful, but removing the new impurity was not. Ultimately, they converted 11 to a tert-butyldimethylsilyl ether, which could be purged by crystallizing 5 from methanol. (Org. Process Res. Dev. DOI: 10.1021/acs.oprd.6b00156)


Remy Angelaud and 11 colleagues at Genentech (South San Francisco, CA) and Siegfried AG (Zovingen, Switzerland) identified the main GTI impurities in the synthesis of vismodegib (12 in Figure 3) as nitrobiaryl starting material 13 and acid chloride intermediate 16. In addition, precursors 1720 to nitrobiaryl 13 were also suspect, as was a hydroxylamine intermediate formed during the catalytic hydrogenation of 13 to 15. (The authors cite two syntheses of 13; 17 and 18 are precursors in one synthesis and 19 and 20 in the other.) 

Figure 3

Based on regulatory guidance, the threshold of toxicological concern of 1.5 g/day of each GTI translated into maximum allowable limits in the final product of <10 ppm for each in a 150-mg daily dose of 12.

The researchers determined the fate of nitro compounds 1720 by spiking 1 wt% (10,000 ppm) of each compound into the hydrogenation reaction of 13 to 15 to assess the purging capability of the process. In all cases, the concentration of each compound was <4 ppm in 13, which translates to a purge factor of >2500. Based on this information, they set the limit of each impurity in 13 at 0.1%.

The investigators then spiked 1 wt% of the hydroxylamine impurity into amine 15 and ran the amide coupling reaction that leads to 12. The hydroxylamine could not be detected in isolated vismodegib, resulting in a purge factor of >10,000. Thus the hydroxylamine limit in amine 15 was set at 0.5%.

Finally, kinetic hydrolysis of acid chloride intermediate 16 at 60 °C predicted a hydrolysis time of 4–15 min for the concentration of 16 in 12 to decrease to 1 ppm. The aqueous workup of crude 12 takes >5 h when the reaction is scaled up, so <1 ppm of 16 remains in 12 after that period. On this basis, it was not necessary to specify a maximum concentration of 16 in 12. (Org. Process Res. Dev. DOI: 10.1021/acs.oprd.6b00208)

Beware the GTIs

These two examples show that many strategies can be taken to reduce impurities, whether they be residual starting materials or reaction byproducts. Extra vigilance and particular care are required if the impurities are MIs/GTIs.