The suspension of clinical trials of a promising new anti-diabetes drug in humans following reports of drug-induced tumours in laboratory mice raises a host of ethical issues.
ETHICS and commercial interests often do not go together. One industrial sector where this conflict is perhaps most evident is pharmaceuticals. It is not uncommon for pharmaceutical companies to tend to exploit regulatory loopholes in a bid to enter the market quickly when the perceived commercial potential for a promising novel drug is big. A case in point is the recent controversial worldwide clinical trials by a multinational company of a new anti-diabetes drug discovered in India.
After several months of Phase III clinical trials of the drug ragaglitazar in 2,550 individuals across 32 countries, the Danish pharmaceutical multinational Novo Nordisk suspended the trials on July 22 after the discovery of urinary bladder tumour in mice on July 8. The drug, codenamed NN622 by the company, was discovered by the Hyderabad-based drug company Dr. Reddy's Laboratories (DRL) and was licensed out to Novo Nordisk in August 1998. Before licensing, the drug was called DRF-2725 by the Indian company and, according to the DRL spokesman, had undergone "preclinical and some preliminary toxicology studies".
The objective of any Phase III trials is to obtain sufficient data about the efficacy and safety of the drug in a large number of patients of both sexes in multiple centres usually in comparison with a standard drug and/or placebo if a standard drug does not exist. On successful completion of Phase III trials, permission is granted to market the drug. In the present case, of the 2,550 persons, about 1,100 had received NN622 while the rest had received another drug, or a placebo, according to a company statement.
Ragaglitazar (NN622/DRF-2725) is a dual-action insulin sensitiser for the treatment of Type 2 diabetes. Insulin sensitisers are substances that stimulate the body's ability to utilise insulin. It is characterised as a "dual-action" drug because of its ability to regulate both sugar and fat levels in blood. In medical jargon, the drug is a "PPAR (peroxisome proliferator-activated receptor) alpha and gamma agonist". It is stated to be chemically and pharmacologically different from currently marketed "PPAR agonist" anti-diabetic drugs and constitutes a new class of insulin sensitisers.
According to media reports, Phase II trials of the drug began in September 2000 and were completed in September 2001, when the company announced that the trials had provided clinical proof of the concept for NN622 and said that it had decided to initiate Phase III trials. Novo Nordisk claimed that in pre-clinical animal tests and Phase I and Phase II human trials, NN622 had demonstrated "significant potential to regulate blood glucose and diabetic dyslipidaemia (or abnormal blood lipid levels)". Phase III trials began in November 2001.
The company claimed that the compound would be among the first to reach the market from a new generation of dual-action sensitisers under development. According to market sources, the nearest competitor to NN622 was Glaxo-SmithKline's Farglitazar (GI262570), which was also withdrawn recently owing to a different kind of adverse reaction. Apparently, the Glaxo drug was ahead of Novo Nordisk in the development cycle. Therefore, NN622 would have been the first to hit the market had Phase III trials reached completion.
The company release of July 22 said: "Novo Nordisk's decision was taken in response to findings of urine bladder tumours in one mouse and a number of rats treated with ragaglitazar. All current clinical trials involving ragaglitazar have been stopped and all planned new clinical trials have been postponed, while preliminary data from studies in rats and mice are being investigated."
Apparently, tumours had been seen in a number of rats in February itself. However, clinical trials were continued because, according to the company, many drugs currently available in the market have been shown to cause tumours in rats and so the findings were not considered to be alarming. A Novo Nordisk release added: "It is a known fact that the development of such tumours in rats is species-specific and does not have any implications for humans. Nevertheless, this data was immediately shared by Novo Nordisk with the regulatory authorities and the trial continued with the knowledge and consent of the regulatory authorities concerned."
According to the company, all the patients were asked to reassess their continued participation in clinical trials, and sign a new Patient Informed Consent Form, which included information about the tumours found in rats. "(Considering) the fact that tumours have now been reported in two species, it was not ethical to continue trials in humans until it can be documented that the mechanism by which these tumours develop is specific to rodents," the release said.
However, the company refrained from disclosing the names of the countries and the centres where the trials had been going on. Details about Indian subjects participating in the trials would perhaps not have come to light but for the fact that, through a PTI story of August 13, Pushpa Bhargava, the eminent biologist and former director of the Centre for Cellular and Molecular Biology (CCMB), Hyderabad, questioned the ethics of carrying out human trials before the animal trials were completed.
As a member of the institutional ethics committee of the Nizam Institute of Medical Sciences (NIMS), Hyderabad, Bhargava had reasons to be concerned as the NIMS was one of the Indian centres participating in the Phase III trials of NN622. According to Novo Nordisk, which responded immediately with a press release, 130 persons from India were part of the suspended NN622 trials and half of them would have received doses of NN622. In response to an e-mailed questionnaire from Frontline, Novo Nordisk provided a partial break-up of the number of patients participating in Phase III trials in country groups rather than individual countries. It also refused to give any more information beyond stating that eight Indian centres were involved in the Phase III trials.
Refuting allegations of illegal and unethical clinical trials, the company stated that the Phase III trials had received approval in all countries in accordance with internationally accepted as well as domestic laws and guidelines governing such trials. "Adequate short-term as well as long-term animal toxicity studies as required including those under Schedule Y of the Indian Drugs and Cosmetics Act, at the current stage of the development of the drug, were provided when the trials were approved," it said.
However, statements made by Novo Nordisk with regard to the nature of these animal trials have been contradictory. While suspending the Phase III trials, the company had described the animal studies as "preclinical" ones, which, by definition, should mean studies conducted before all phases of human trials. "Preclinical carcinogenicity studies," the July 22 release said, "are an integral part of the development of new drugs for chronic use. The purpose of doing preclinical carcinogenicity studies in animals is to investigate if a new drug has potential for causing tumours after long exposure in animals. Due to complexity of conducting state-of-the-art preclinical carcinogenicity studies data from such studies do not usually become available until very late in Phase III clinical development." The last sentence would seem self-contradictory.
However, an August 14 company release, issued in response to news reports in the Indian media, stated that the mouse that had developed tumour had been treated with ragaglitazar for almost two years. Given that Phase II trials had begun in September 2000, it is clear that the animal tests could not have been "preclinical", even assuming that the term is meant to indicate only the starting date, as Phase I trials should have lasted at least six to eight months. So in all likelihood, the long-term carcinogenicity studies (in rats and mice) were initiated well after the human trials had started. (The company did not provide dates and other details of Phase I and Phase II trials that were sought through an e-mail query.)
The apparent contradiction may seem somewhat academic because, in any case, the results of animal trials were not available before the human trials. But therein lies the crux of the matter, an issue of ethics - whether human trials should be initiated before results of animal trials relating to long-term carcinogenicity become available.
Countering criticism that the animal trials were not initiated well in advance, the release further said: "Some types of animal studies must be performed before starting human trials... These studies show whether the new medication has any unwarranted side effects which develop in less than six months of treatment. Safety data from six-month toxicity studies on two animal species (rat and dog) were made available to the authorities as required by law, before the human studies were initiated...The carcinogenicity studies must be performed on compounds that are developed where treatment of patients is expected to exceed more than six months. According to international guidelines, these types of long-term carcinogenicity studies need not be conducted prior to entering Phase III development. The study data is only required at the time of (marketing) approval of the medicine."
Said Bhargava: "I do not care about laws and regulations. It is simply unethical to begin human trials before animal tests are over. I can understand (this argument) in the case of HIV or cancer where there is no treatment available and it is in human interest to reach a new potential drug quickly to the people. But diabetes is not incurable and there are a whole lot of drugs."
"The existing system of guidelines for clinical trials and long-term carcinogenicity studies do not address the case of drug development for chronic use adequately. Even the guidelines of the Indian Council of Medical Research (ICMR) lack clarity in this respect," points out Vineeta Bal of the National Institute of Immunology (NII), New Delhi, who has been concerned with ethical issues in biomedical research.
Vineeta Bal added: "It is precisely for drugs meant for chronic use that long-term carcinogenicity results should be known before you begin clinical trials. Generally, such animal studies will take about three to four years. Companies are not willing to wait that long. So they carry out such studies in parallel with clinical trials, which is not ethical. There could be some flexibility in the case of a disease like HIV. But this flexibility has to be made more specific instead of the existing vagueness, which companies exploit to their advantage."
Said Vasantha Muthuswamy of the ICMR: "We insist on animal test results as per the requirements of the Schedule Y of the Indian Drugs and Cosmetics Act to be available before clinical trials can be approved whenever a drug is referred to us. While the referral process is meant to be mandatory, some cases are sent to us and some are not. This particular case did not come to us." However, there is inherent vagueness and an apparent inconsistency in Schedule Y with regard to carcinogenicity studies. As a result, it is open to interpretation and it may be argued - as indeed Novo Nordisk has - that the company had complied with the requirements under Schedule Y.
A CLOSER reading of the Schedule shows that Novo Nordisk has not complied with other aspects of the regulations as well. In Schedule Y, under Animal Toxicity (Appendix I, Item 4), a clear distinction has been made between long-term animal toxicity data and mutagenicity/carcinogenicity data that are required to be submitted for obtaining clearance for clinical trials (Article 1.2). Long-term toxicity studies are required to be carried out in at least two mammalian species, one of which should be non-rodent. The Schedule also specifies (Appendix III) that for drugs to be administered (orally, parenterally or transdermally) to humans for a period longer than 3 months, animal test data (in the two species) over a period of six months are required to be submitted if the clearance required is for Phase III and/or marketing permission (MP).
As the press release of August 14 has stated, it is this 6-month data (in rat and dog) that the company had supplied to the Drug Controller General of India (DCGI) in order to get approval for clinical trials under the Indian law. The response of Anil Kapur, Managing Director of Novo Nordisk India Pvt. Ltd, to Frontline's e-mail questionnaire makes it clear that the term long-term toxicity has been interpreted to include carcinogenicity as well. If it was a question of semantics alone, it could be argued that carcinogenicity is a facet of long-term toxicity and, therefore, requirements under Schedule Y have been met. But the requirements of mutagenicity/carcinogenicity studies data are different and have been specified in Article 3.5 of Schedule Y. It states: "These studies are required to be carried out if the drug or its metabolite is related to a known carcinogen or when the nature and action of the drug is such as to suggest a carcinogenic/mutagenic potential. For carcinogenicity studies, at least two species should be used. These species should not have a high incidence of spontaneous tumours and should preferably be known to metabolise the drug in the same manner as humans... The drug should be administered 7 days a week for a fraction of the lifespan comparable to the fraction of human lifespan over which the drug is likely to be administered therapeutically...etc. etc."
According to a fact sheet on NN622 issued by Novo Nordisk on July 26, all genotoxic tests on the compound were negative, indicating that the drug in all likelihood does not have any genotoxic carcinogenicity potential. A genotoxic compound affects the genetic material. A non-genotoxic carcinogen requires a long exposure period, and when a drug is meant for chronic use, its potential non-genotoxic carcinogenicity needs to be ruled out. The long-term carcinogenicity studies on NN622 were being carried out from this perspective. In the Indian context, the very fact that such studies were being conducted means that they should conform to Article 3.5 of Schedule Y.
Novo Nordisk has violated this requirement on two counts. One involves the choice of rat as one of the species, when according to the company's own statements, spontaneous tumours are known to arise in rats with many of the currently marketed drugs. Also, in response to the Frontline questionnaire, Kapur stated that a similar drug that has been marketed had resulted in tumours in rats. Given this situation, conforming to Schedule Y would call for the choice of a species other than rats. Two, the average life-span of bred mice is about two to three years. Assuming that the average period over which a Type 2 diabetes drug will be administered would be half the human lifespan, the carcinogenicity data needed for permission for clinical trials should span at least a year, and not six months as in the case of the data submitted by the company.
The company has claimed that the tumours occurred after the mice were exposed for a period of time equivalent to between half and full life expectancy. The patients, on the other hand, were exposed to a maximum of six to seven months, equivalent to about 0.5-1 per cent of human life-span. In the Indian case, since the Phase III trials had begun only in March, the period of exposure was only four months. Therefore, according to the company, the risks for the human subjects of trials are minimal.
The issue is not so much the risk of exposure but the ethics of conducting trials by the company. "You cannot proceed with a general assumption that tumours seen in rats are spontaneous," points out Bal. "Spontaneous tumours are dependent on the class of drugs being tested, dosage and several other factors. We have seen spontaneous tumours in dogs for certain drugs."
Clearly, the company cannot have it both ways. If it was known that spontaneous rat tumours are likely, the species should not have been used at all. If not, ethics demanded that the cause of rat tumours be established - determine whether they were indeed spontaneous tumours - before continuing with the trials.
More fundamentally, the company would seem to have not fulfilled even the basic requirement for conducting Phase III trials under Schedule Y. The Schedule says: "If the drug is a new drug substance discovered in India and not marketed in any other country, Phase III data should be obtained on at least 500 persons distributed over 10 to 15 centres." By choosing to have only 130 patients, distributed across eight centres, Novo Nordisk violated this regulation.
Legally Novo Nordisk cannot be questioned because it had obtained the DCGI's clearance for its clinical trials. The DCGI has, however, refused to answer Frontline's questions on the subject. ICMR sources say that when the council asked the DCGI why the application was not referred to the ICMR and under what circumstances the approval was given, the DCGI stated that the clearance was given since this was an international study and involved multi-country trials. In particular, the DCGI would have to explain why trials were allowed even after it was informed of rat tumours in February.
According to Novo Nordisk, the risk of exposure to humans participating in trials was negligible on two counts. First, the duration of exposure was much less than the period over which the animals were exposed. Second, the tumours may be rodent-specific. Novo Nordisk has argued that rodent urine is significantly different in composition from human urine and it is possible that tumours were caused by a mechanism leading to the formation of crystals in the rodent urine. The formation of such macro- or micro-crystals might eventually be found to be the cause of the carcinogenicity findings in the urinary tract system in the rodent, in which case, the mechanism would be of no relevance to humans, the company has stated.
In any case, the company's fact sheet has stated that all patients will be called for an end-of-trial visit. The company has suggested that patients have a urine sample taken a year after the drug trials in order to document any adverse drug reaction (ADR). However, in response to an e-mail question, the company refused to provide details of the in-built liability clause in the Patient Consent Form and whether the company would bear the costs of treatment if any ADR linked to NN622 were to be detected. It merely said: "Novo Nordisk will work according to the liability clauses valid in the individual protocols and patient consent forms."
THE Swiss company Novartis Pharma AG had entered into an agreement with Novo Nordisk in July 2001 for obtaining commercial rights over NN622 in the United States, Canada and Mexico against some milestone payments to the latter. However, Novartis terminated the agreement in October 2001. There has been speculation in pharma business circles that perhaps Novartis had some inkling early on of the problems relating to animal tests.
The DRL-Novo Nordisk licensing agreement was the first instance of an Indian drug discovery being licensed to a foreign multinational for the international market. It was hailed as an indication of the emerging Indian R&D potential in the ongoing globalisation in the pharma sector. The setback may dent that image somewhat but does not lower that potential. What the issue has brought to light is how ethics are sidestepped in favour of overriding commercial interests and how the regulatory mechanisms of the country are becoming weaker even as MNCs enter the Indian market in a big way. In this context, it would be unwise to implement the recent recommendation of the Planning Commission Working Group to ease regulations on clinical trials and revise Schedule Y accordingly without a thorough revamping of the functioning of the DCGI itself.
Break up of patients in Phase III trials of NN622 in country-groups:
United States/Canada - 650 Latin America - 200 Australia/New Zealand - 100 European Union countries (Scandinavia, United Kingdom, France, Italy, Belgium, Austria, Germany, Ireland, The Netherlands) - 800 Non-E.U. European countries (Estonia, Switzerland, Hungary, Poland, Slovenia) - 250 Asia (India, Hong Kong, Malaysia, Singapore, Philippines, Taiwan, Thailand) - 550; (130 in India alone spread over eight centres.)