IAS: Roche admits long-standing low-level contamination of nelfinavir

This article is more than 13 years old. Click here for more recent articles on this topic

Nelfinavir (Viracept) manufacturer, Roche, has admitted that some batches of the now-withdrawn protease inhibitor have been contaminated with unacceptable levels of a cancer-causing substance since production began in 1998, although the very highest levels of contaminated products did not reach patients until March 2007. The United Kingdom is just one of twenty countries worldwide where batches containing the highest levels of contaminated drug have been distributed, and Roche is currently establishing two global patient registries in order to monitor for possible long-term effects.

Roche provided many further details regarding the contamination and near-global recall of the drug in a meeting with community treatment advocates that took place in Sydney during the 4th IAS conference.

Batches of nelfinavir contaminated since production began

Dr Malte Schutz, introduced at the meeting as “Roche’s clinical expert”, said that Roche now have much more information on the level of ethyl methane sulphonate (EMS), the contaminant that is known to react with DNA to cause genetic changes that may result in cancer or birth defects, although current data are based on animal studies.

Roche, he said, has now tested “all batches, going back to the time of [EU] marketing authorisation”, which took place in January 1998. Although the highest level of EMS (2,300 parts per million, or ppm) reached patients after March 2007, some batches produced between 1998 and 2003 contained EMS at levels up to 25 ppm, and some of those produced between January 2004 and March 2007 contained levels up to 132 ppm.


deoxyribonucleic acid (DNA)

The material in the nucleus of a cell where genetic information is stored.

concentration (of a drug)

The level of a drug in the blood or other body fluid or tissue.


An unborn baby.

European Medicines Agency (EMA)

Regulatory agency that evaluates medicines for safety and efficacy in Europe, performing a similar role to the Food and Drug Administration (FDA) in the United States. The EMA recommends to the European Commission that a medicine can be marketed in the European Union and European Economic Area.

Countries directly affected by the highest level of contaminant (over 1000 ppm) are:

Botswana, Burkina-Faso, Cameroon, Egypt, France, Germany, Iran, Italy, Kenya, Mali, Mexico, Mozambique, Nigeria, Portugal, South Africa, Spain, Taiwan, Uganda, Ukraine, and the UK.

Dr Schutz told the meeting that “EMS is not an agent that is routinely tested for,” even though, in 2001, a risk assessment determined that EMS may occur during the production of nelfinavir. In 2001, the European Medicines Agency (EMEA) asked Roche to test batches of the drug for EMS, and at the time the batches tested contained what the EMEA considered a safe level of EMS (less than 1 ppm). He said that the EMEA did not require Roche to test the product further, and until the ‘strange smell’ alert from patients in June 2007, they had not done so.

He added that no official guidance on acceptable EMS levels was produced by the EMEA until 2005. They have now agreed a lower level (0.6 ppm) for all future production of nelfinavir. During the question and answer session, Dr Schutz said that during their meetings with the EMEA it was decided that the risk for adults ingesting nelfinavir that may have contained EMS below 1000 ppm was “extremely low” and that close follow-up was not necessary.

However, individuals from the 20 countries listed above who received nelfinavir after March 1st are considered to have been potentially exposed to EMS over 1000 ppm, and will make up one of two global registries, details of which are being finalised. The primary outcome of this registry, said Dr Schutz is to “find out if any harm was done in terms of increase of cancer rates.”

The second registry will include all women who have ever received nelfinavir while pregnant as well their children exposed in the womb, and all children (

Understanding the risk from EMS ingestion

Dr Schutz noted that the contaminant “can be the cause of future cancer development and that is clearly the key concern about EMS.” However, he said that research published earlier this year suggests a threshold effect. “If you have low exposure to EMS,” said Dr Schutz, “the DNA damage appears to be repaired straight away.”

He explained that they had also discovered that an individual’s highest daily exposure risk – thought to be 2300 ppm in batches produced after March 2007 – was, in fact, lower following results of an analysis done after June 6th of this year of Viracept tablets. This analysis, he said, found that EMS levels decreased by 60% when the nelfinavir powder that comes off the production line is generated into tablets. “So, the highest concentration in tablets is actually 960ppm,” he concluded.

Previous calculations on cancer risks, based on animal studies, suggest that even at the highest levels, exposure is around 200-fold lower than was seen to cause cancer in mice and rats.

Based on another study, examining the DNA of mice embryos exposed to EMS in the womb, the theoretical risk of genetic abnormalities resulting from a mother ingesting contaminated nelfinavir during pregnancy was “1 in 20,000,” said Dr Schutz. “Although that would not be a risk that you would want to have,” he continued, “the normal malformation rate in humans is 3%, and this risk is 0.005%.” Consequently, he said, “we expect the risk of damage to a foetus to be extremely low”.

However, noted Dr Schutz, there are many limitations to all of the studies done so far on EMS. Consequently, the EMEA has asked Roche to undertake “more robust studies” which will begin next month.

These will again be done in mice – because the EMEA does not consider EMS studies in primates (or humans) to be ethical – and will have doses based on likely human exposure to EMS at the highest levels seen in the most recent batches of nelfinavir for between one and three months.

The first will look at genetic mutations and the second will examine chromosomal damage to DNA. Results are expected in December 2007, with Roche promising an interim update in October 2007.

Dr Schutz added that the “animal data will allow us to revisit the necessity and scope of the [two patient] registries.”

What went wrong with the manufacturing?

Ethanol had been routinely used to clean the tank holding the raw material for nelfinavir – methane sulfonic acid – at Roche’s Swiss manufacturing plant since production began in 1998.

The tank was last cleaned at the end of 2006. “We know now,” said Dr Schutz, “that the tank wasn’t completely dried out, and there was about two litres left – but it’s a large tank.” When the methane sulfonic acid was added to the tank at the start of production in January 2007, this resulted in “a slow chemical reaction.”

The first three batches of nelfinavir produced in 2007 contained three, five and seven parts per million of EMS. “Then,” explained Dr Schutz, “the production line sat idle for 77 days. This had never happened before.” During the question and answer session that followed, Dr Schutz said that “production is driven by demand, and there was not as much demand [for nelfinavir] as there used to be.” It was this chain of events that led some of the March 2007 batches to be contaminated with the highest levels of EMS.

Legal and financial liability issues

“What we are willing to pay right now,” Ms Edge-Dallas told the meeting “is compensation around the recall of Viracept. We cannot speculate any more at this time, until we know in actual fact whether there is any damage caused by EMS.

Roche had since clarified their position in a press release with the headline, ‘Roche inaccurately cited as offering donation of replacement drug’.

“Roche will reimburse all reasonable costs and expenses related to the recall,” it says, “such as additional doctor visits and costs linked to the registration into the patient registry."

“However, the provision of replacement therapy is a different issue. The nature of HIV means that treatments need to be tailored for the individual patient. With complex and complicated treatment regimens, even in resource-limited settings, it is not ethically or medically possible for Roche to provide just one option for patients to switch to."

“Prior treatments, virological factors, CD4 count, prior medication and polypharmacy make it essential that Viracept patients seek advice and consultation with their treatment specialists in order to switch."

“Roche takes this very seriously,” said Ms Edge-Dallas , “and our biggest concern is for the patient.”

Future availability and other issues in developing countries

Another Roche spokesperson, Maria Vigneau, responded to criticism from some of the treatment advocates at the meeting that the company did not do enough – or work fast enough – to recall the drug and offer alternatives in resource-limited countries. “We do accept that we cannot possibly know everyone who is provided treatment,” she said. “Only now we are finding that Viracept is being sold on to other treatment providers in these countries. We know who has bought [drug] from Roche Basel. If there’s additional information, of people purchasing [nelfinavir] via other routes, then we would appreciate the information.”

She added that Roche is working with their African distributors to identify clinics supplied with nelfinavir and that they are providing “all the info we possibly can.”

Roche has since announced that a meeting with NGOs and key stakeholders is scheduled to take place on August 30th in Geneva “to work in partnership on local challenges across Africa to collect patient registry information.”

Following the EMEA’s suspension of nelfinavir’s licence on June 20th, Roche went back to the EMEA in response to requests from patients, clinicians and advocates to try and resupply uncontaminated nelfinavir where it was most desperately needed. However, Jenny Edge-Dallas, Roche’s HIV franchise global leader, told the meeting that “the EMEA turned down the application.”

She added that they continue to hear of demand for the drug, and suggested to the assembled treatment advocates that it would be helpful to Roche “if you could write letters...to help us [get the uncontaminated batches of nelfinavir] released. We have drug available with less than 0.6 ppm [of EMS], but the EMEA will not allow us to release it.”

She also said that Roche is “talking to Pfizer (which manufactures nelfinavir in the United States for the US, Canada and Japan), but this is not a process that happens very, very quickly. We have tried very hard to have some outcome in the next weeks as to whether it is going to be possible or not, but I can’t give you a definitive date.”

References, and further information

For more information on the nelfinavir recall, including the Powerpoint presentation shown at the IAS meeting; an overview for resource-limited countries; further information on the patients registries and patient advice is available on the Roche-HIV website.