Launch Conference Final version: ECA's Integrated Qualification and Validation Guide

Working with Suppliers towards modern Qualification and Validation

27/28 October 2020 in Berlin, Germany (on site) or broadcasted live to your desk!

Novartis's drug patent for valsartan expired in 2011 in major European countries, and in September 2012 in the United States1. This context might be key to understand the contamination event of Zhejiang Huahai Pharmaceutical Company's valsartan product with N-nitroso dimethylamine (NDMA), a probable carcinogenic impurity2. As that date was approaching, manufacturers of generic active pharmaceutical ingredients (API) were racing to develop their own synthesis processes of this substance. A market that in 2010 was worth six billion USD was about to open its doors1.

Judging from the patents fi led around that time, much of the attention of the generic manufacturers focused on the last step of synthesis of the valsartan molecule: the formation of the 5-substituted tetrazole cycle. Novartis/Ciba- Geigy's original method for the tetrazole-forming step relied on the use of tributyltin azide. The yield of this step was 65%. Such a poor performance might have been good enough as long as the patent was keeping competitors at bay, but not good enough for the manufacturers of generic pharmaceutical ingredients racing for the incoming market shares. Higher synthesis yields were critical to achieve prices low enough to capture the business of the pharma customer companies.

Various methods for forming 5-substituted tetrazole cycles on nitrile groups with sodium azide had been around for some time, in presence of metal catalysts, strong Lewis acids or tertiary amines3. In September 2010 Zhejiang Menovo patented a sodium azide-based synthesis method for valsartan, resulting in much improved yields approaching 90%4. The Menovo inventors boasted: "The method has the advantages that the operation is simple; the yield is high; the product purity is high; and the industrial production is easy"5.

Considering such advantages, it was hardly surprising that Zhejiang Huahai Pharmaceutical Company was developing its own process in a similar direction. There was a difficulty, however. Sodium azide was a dangerous substance whose toxicity was comparable to that of potassium cyanide. It was fatal if swallowed, inhaled or in contact with the skin6. Since the new synthesis method required bringing a stoichiometric excess of sodium azide to the reaction, an unreacted residue of this highly toxic substance would remain in the product. This issue had to be dealt with.

Aware, as any medicinal chemist would be, of the toxicity of sodium azide, Zhejiang Huahai's valsartan process chemists opted for adding nitrous acid to the production broth as soon as an in-process control would confirm the complete formation of the tetrazole cycle. They did so in the form of sodium nitrite, which in the acidic environment of the broth turned into nitrous acid7,8,9. Nitrous acid was a known decontaminating agent of sodium azide. It caused the degradation of sodium azide to nitrogen gas, nitric oxide and sodium hydroxide10. The toxicity issue was thus resolved. Zhejiang Huahai patented this production method in 20149, but their valsartan process had been based on the use of sodium azide as far back as 20127,11. By 2017 their annual sales of valsartan had grown to 328 million Yuan (41 million Euro)12.

Tragically, Zhejiang Huahai's chemists failed to realize that while eliminating the toxicity due to sodium azide, their process generated another toxic substance: N-nitroso dimethylamine (NDMA), a probable carcinogen in humans13. The solvent of the tetrazole-forming step was dimethylformamide, whose industrial production process started from dimethylamine14. It is now suspected that the residue of dimethylamine in the solvent reacted with nitrous acid to generate N-nitroso dimethylamine7,8. The presence of this contaminant remained undetected until early June 2018, as the manufacturer was reviewing and optimizing its processes15.

It is important to underline that the levels of NDMA eventually identified in valsartan were well below 0.05%, the disregard limit of the European and American pharmacopoeias for the related substances of valsartan16,17. According to the European Medicines Agency (EMA), NDMA was found in Zhejiang Huahai's valsartan, on average at 66.5 ppm, and at most at 120 ppm - enough to represent an unacceptable health hazard, but not enough to be distinguished from analytical noise during routine quality control of valsartan8. Besides, whereas HPLC was used for control of the related substances of valsartan, gas chromatography would be the preferred method for separating NDMA18. In 2017 FDA investigators had reported occasional issues with the tests of impurities during inspections of Zhejiang Huahai, but those issues were not critical and concerned chromatography peaks above the disregard limit19. Since it was below the normal reporting level of impurities, the presence of NDMA had been missed for years, not due to any default of the company's quality control, but as a consequence of the faulty assessment of the safety of the synthesis process.

Nor was it solely Zhejiang Huahai’s failure. The presence of the NDMA contaminant implied that the company's certificate of suitability of the European pharmacopoeia (CEP) for valsartan was wrong20. The European Directorate for the Quality of Medicines (EDQM) suspended the validity of the certificate as soon as the contamination became known21. The EDQM assessors of valsartan's CEP application dossier had failed to realize the risk - as had Zhejiang Huahai, the US FDA and everyone else. The risk of formation of carcinogenic nitrosamines from nitrites and secondary amines in acidic conditions has long been known to the food industry. It was not rocket science, yet it came as a surprise to the pharmaceutical industry. The use of sodium nitrite in the valsartan processes of Zhejiang Huahai and other manufacturers should have been a red flag prompting an evaluation of any presence of secondary amines, but it did not.

In this perspective, it would be a mistake to see Zhejiang Huahai as an exceptionally careless actor. In fact, quite the contrary - and there is an important point to make about it. This company actually managed the risk of genotoxic and mutagenic impurities better than the average manufacturer. For instance, in one of their products, dabigatran etexilate mesilate, Zhejiang Huahai had specified a tight ppmlevel acceptance limit for control of the potential presence of ethyl mesilate, a genotoxic substance known for having caused a carcinogenic contamination in nelfinavir at a Swiss plant in 200722. Such careful control is not commonly seen at the manufacturing plants of mesilate substances. Zhejiang Huahai's discovery of the presence of NDMA in valsartan in June 2018 was of course overdue, but anyone familiar with the API industry will recognize that it was beyond what the average generic API company would care doing. How often do we see an effective ICH M7 compliance program at any API plant? Rather than any exceptional sin on the part of this manufacturer, the valsartan contamination event might reveal that the pharmaceutical industry, focused as it is on the quality of the pharmaceutical ingredients, has not been paying enough attention to the safety of the chemical synthesis processes.

As soon as the valsartan contamination became known, the manufacturer and regulatory authorities rushed to test NDMA in the other products of the company2. They were satisfied to detect no NDMA in those products. Irbesartan, which is another tetrazole molecule, was one of those products. Like valsartan, Zhejiang Huahai synthesized it according to the sodium azide method. Contrary to valsartan, though, no dimethylformamide was used - but triethylamine was, which implied the natural presence of a residue of diethylamine from the manufacturing process of triethylamine by alkylation of ammonia with ethanol23,24. Hence, in case sodium nitrite was used in this process too - Zhejiang Huahai said it was not when I asked them on 15 August 2018 - the potential contaminant that needed to be tested in irbesartan was, not N-nitroso dimethylamine (NDMA), but Nnitroso diethylamine (NDEA). Everyone was looking for the wrong impurity. Just like the valsartan process, the chemical synthesis process of irbesartan had not been carefully looked at.

The scope of investigation was extended to other valsartan manufacturers, too. On 16 July 2018, the European Medicines Agency wrote to all manufacturers of this substance to inquire about their production methods for a first evaluation of risk of generation of NDMA8. Obviously, the investigation will also need to be extended to candesartan cilexetil, irbesartan, losartan and olmesartan medoxomil, too, as these five substances contain 5-substituted tetrazole cycles that are formed through to the same method based on sodium azide23,25. It seems we will have to look deeper and wider than just NDMA in valsartan.

We can be confident the valsartan contamination incident will be corrected - although some patients will suffer serious consequences. Zhejiang Huahai reacted to their discovery responsibly. They halted production, sealed the stocks at the warehouse, and promptly notified the authorities and customers7. The synthesis process will be modified, and NDMA will be routinely evaluated in valsartan7,18. This event will eventually be brought under control. But we would do well to understand that the contamination of valsartan remained undetected for years because the industry's approach to the quality control of pharmaceutical ingredients rather focuses on the generally harmless related substances of synthesis.

The troubling question is: How many NDMA, NDEA, ethyl mesilate and other highly toxic, low-concentration, impurities might lie hidden in the hundreds of pharmaceutical ingredients made by the thousands of manufacturers of the world? For instance, who checks the purge of 1-chloro-4-nitrobenzene in the synthesis process of paracetamol? The purge of benzyl chloride in the synthesis process of dapaglifozine? How many such hazards might be there under the radar? Something needs to change in our quality management of pharmaceutical ingredients, not only to prevent similar contaminants in the future, but also to become aware of the ones that may be there, right now, in our medicines.

Philippe André
... is GMP Auditor.

1 Novartis Group, Annual Report of 2010
2 European Medicines Agency, 5 July 2018: "EMA reviewing medicines containing valsartan from Zhejiang Huahai following detection of an impurity" (EMA/459276/2018) (retrieved on 2 September 2018)
3 Organic Chemistry Portal: "Synthesis of Tetrazoles" (retrieved on 19 August 2018)
4 Weiren Chen, Doxin Gong, Zhejiang Menovo Pharmaceutical Company, 1 September 2010: Patent CN101817795 ?????????????? "Improved method for synthesizing valsartan"
5 Europe PMC Repository: (retrieved on 19 August 2018)
6 European Chemicals Agency: "Sodium Azide Substance Information" (retrieved on 19 August 2018)
7 Zhejiang Huahai Pharmaceutical Company, 10 August 2018: "NDMA Impurity in Valsartan Q&A" (retrieved on 8 September 2018)
8 European Medicines Agency, 16 July 2018 "CHMP List of questions" (EMA/CHMP/467845/2018) (retrieved on 8 September 2018)
9 Xiaoren Zhu, Nianping Shan, Wenling Zhang, Peng Wang, Zhejiang Huahai Pharmaceutical Company, 17 September 2014: Patent CN1040045602A 《一种缬沙坦成四氮唑的改进方法》 "Improved method for forming the tetrazole cycle of valsartan"
10 US National Academy of Sciences, National Research Council, 1995: "Prudent Practices in the Laboratory: Handling and Disposal of Chemicals, Revised Edition" (ISBN 0-309-05229-7)
11 Zhejiang Huahai Pharmaceutical Company, May 2012: "Route of Synthesis for Valsartan"
12 South China Morning Post, 18 July 2018: "Chinese blood pressure pills sold in US recalled over cancer-linked ingredient" (retrieved on 8 September 2018)
13 International Conference on Harmonization: Addendum to ICH M7(R1), step 4 version of 31 March 2017: "Assessment and Control of DNA Reactive (Mutagenic) Impurities in Pharmaceuticals to Limit Potential Carcinogenic Risk" (retrieved on 8 September 2018)
14 European Chemicals Agency, 6 February 2014: "Background document for N,N-Dimethylformamide" (retrieved on 31 August 2018)
15 Zhejiang Huahai Pharmaceutical Company, 21 August 2018: "Update on Valsartan API - A Statement From the Company" (retrieved on 5 September 2018)
16 The European pharmacopoeia: "Valsartan" monograph 01/2010:2423
17 The United States Pharmacopoeia, 41st edition: "Valsartan" monograph
18 US Food and Drug Administration, 30 August 2018: "Statement from FDA Commissioner Scott Gottlieb, M.D., and Janet Woodcock, M.D., director of the Center for Drug Evaluation and Research on FDA's ongoing investigation into valsartan impurities and recalls and an update on FDA's current fi ndings" (retrieved on 3 September 2018)
19 US Food and Drug Administration: Form FDA-483 of 19 May 2017 following inspection of Zhejiang Huahai Pharmaceutical Co., Ltd. on 15-19 May 2017
20 The European Directorate for the Quality of Medicines and Healthcare (EDQM), 9 June 2016: Certifi cate of Suitability R1-CEP 2010-072-Rev 00
21 The European Directorate for the Quality of Medicines and Healthcare (EDQM): Certifi cation Database of Suitability of the European Pharmacopoeia https:// DateBtwBegin=&vtsubDateBtwEnd=&SWTP=1&OK=Search (retrieved on 25 August 2018)
22 European Medicines Agency, 6 June 2007: "Press Release European Medicines Agency announces recall of Viracept" (retrieved 30 August 2018)
23 Zhejiang Huahai Pharmaceutical Company: "Route of Synthesis for Irbesartan" (undated document distributed on 14 August 2017)
24 K.S.Hayes, Applied Catalysis A: General, Volume 221, Issues 1-2, 30 November 2001, Pages 187-195: "Industrial Processes for Manufacturing Amines"
25 The European Directorate for the Quality of Medicines and Healthcare (EDQM), 28 August 2018: "Update on EDQM's actions following detection of impurity in valsartan" (retrieved on 2 September 2018)

Go back