As the Covid-19 pandemic spread across the world two years ago, one of India’s leading biotech companies was racing to develop a vaccine with crucial backing from the Indian government. The shot engineered by Bharat Biotech was, in part, an important effort to create a home-grown product that could bolster the fortunes of the Indian pharmaceutical industry.
However, a STAT review of documents detailing the steps taken toward government approval found that regulators endorsed the vaccine, called Covaxin, despite discrepancies in the number of clinical trial participants. Moreover, questionable changes were made to the trial protocols — which are established procedures for testing a vaccine or medicine — to expedite the approval process.
For instance, the number of people enrolled in the Phase 1 portion of the trial differed from what was later published in a medical journal. There were also important changes made to the protocol for Phase 2 testing, when immunogenicity data from the previous Phase 1 stage were not yet available.
In addition, the protocol for Phase 3 was approved while Phase 2 was still underway and the final vaccine candidate was selected without Phase 2 data, according to protocol documents and minutes of meetings held by an expert committee that reported to India’s Central Drugs Standard Control Organization (CDSCO), the national regulator responsible for approving medicines. This was the agency that authorized the vaccine for emergency use in January 2021, two months before Phase 3 results were known.
Consequently, experts in public health and clinical trial design expressed concern that the Bharat Biotech vaccine may have been hastily approved. In the process, they suggested the episode may rekindle questions about the commitment by Indian government agencies to ensuring high-quality medical products are reaching patients.
“My issue is that the confidence of the public and medical community in the regulator takes a beating when the regulator and the (expert committee) seem to accept ‘inadequate information’ to base their decisions on — or lack the clarity in their minutes to explain why this was rational. They appear to be under duress and this is a very worrying scenario,” said John Jacob, a community health professor at Christian Medical College in Vellore, India, who reviewed the documents.
There are circumstances that permit such protocol changes, such as seamless adaptive trials, which combine different phases and allow a trial design to be modified when study is underway. But the parameters for potential changes should first be defined in a protocol, which did not always occur with the Covaxin approval process, and there were no scientific justifications offered for the changes.
For their part, Bharat Biotech executives acknowledged some mistakes, but maintained their decisions reflected the unusual challenges of working amid the pandemic. They also argued they faced “political” pressure to get a vaccine out of the laboratory door as quickly as possible, but denied taking any shortcuts. And they insisted the steps taken to speed the trial were vetted during discussions with regulators.
“In a classic sense of product development, we would do everything the right way — play by the book and all the rules of the game would be followed. But here was a situation the world didn’t foresee,” Krishna Mohan, a Bharat Biotech director, told us. “… Please don’t think there was any issue with the veracity of the data. Yes, it was an unusual approach, but it was dictated by the nature of the pandemic.”
The effort to bring Covaxin to market reflected a convergence of several key concerns.
The Bharat Biotech vaccine was being developed at a time when options were increasingly needed to combat the pandemic. The shot was viewed desirably by the World Health Organization because, unlike the mRNA vaccines from Moderna and Pfizer/BioNTech, the Covaxin shot does not require storage at very low temperatures, which made it easier to deploy in low- and middle-income countries.
Moreover, the Indian government was banking on a locally made vaccine to fulfill national health demands, which helps explain why the Indian Council of Medical Research, a government agency, collaborated with Bharat Biotech to develop Covaxin. That had the added benefit of burnishing the image of the domestic pharmaceutical industry, which has regularly been criticized by overseas regulators.
But the Indian government, which collects a 5% royalty on sales and gave the company a $200 million manufacturing grant, has also been accused of lax oversight. The point has often been underscored by a steady stream of warnings and notices issued in recent years by the U.S. Food and Drug Administration following inspections of manufacturing facilities run by the large number of drugmakers based in the country. In doing so, the FDA has made Indian regulators look slipshod.
“The government of India is in denial about the quality crisis facing the pharmaceutical industry,” said Dinesh Thakur, a whistleblower who disclosed that Ranbaxy Laboratories — now owned by Sun Pharmaceutical — falsified testing data and manufactured drugs that failed to meet safety standards. The company pleaded guilty to U.S. felony charges and paid $500 million to settle fines and civil claims.
“As far as (the government) is concerned, it wants to see the Indian pharmaceutical industry grow, often at the expense of public health. It will not put in place or enforce any regulations that threaten the growth of the … industry,” said Thakur, whose foundation made grants to a journalist at an Indian website who was recently targeted by Bharat Biotech with a court injunction for stories about its shot.
From the start, a lack of transparency surrounded the effort to develop the vaccine.
For months, the documents about the deal between the Indian Council on Medical Research and Bharat Biotech, as well as the Phase 1/2 and Phase 3 protocols were not made publicly available. And during Phase 1, the company acknowledged a serious adverse event in a patient, which it later determined was not tied to the shot. But the disclosure came only after the episode was disclosed by the media.
This was not in keeping with informed consent requirements, according to Malini Aisola and Siddhartha Das, civil society activists who specialize in public health issues and investigated the Covaxin approval process, which included raising questions about irregularities in the early-stage trials. They also worked with the All India Drug Action Network to persuade the CDSCO to eventually release the minutes of expert committee meetings in which Covid-19 vaccines were reviewed.
“Protocols were not in public domain while the trials were happening, and the trials were happening at breakneck speed with no information about the basis for granting approvals,” said Aisola, adding that the handling of the adverse event also set a poor tone for investigating and sharing information with the public. “This isn’t just a purely scientific issue. It involves the role of the government and its partner, which was more than happy to cooperate, because it benefited as well.”
More controversy erupted last spring. Brazilian authorities raised concerns about Bharat Biotech manufacturing. Then, the WHO, which listed the vaccine for emergency use in November 2021, suspended supplies after an inspection of the facilities found unspecified problems. The decision meant United Nations procurement agencies, such as UNICEF, would no longer be able to supply the shot to other countries. A WHO spokesperson declined to offer an update on the findings.
For now, it remains unclear whether the newly disclosed issues surrounding the clinical trial will trigger still more questions about the willingness of the Indian government to boost its oversight. The CDSCO and the Drugs Controller General of India, which oversees the CDSCO, did not respond to emails seeking comment about the changes made to the Covaxin trial protocols and subsequent government approval.
In reviewing the documents, there was a clear discrepancy in the number of enrollees. In reporting the Phase 1/2 data, the protocol stated 402 participants were given the first dose and 394 got the second dose. But results published in Lancet Infectious Diseases in January 2021 stated 375 people were given a first dose and 368 received a second dose. (See Figure 1 on page 640.)
Mohan conceded the error. But he explained that the company immediately encountered difficulties because a large amount of work had to be done remotely, which complicated communications among three different teams of staffers handling clinical operations, data, and interactions with regulators. This also made it difficult at times to communicate with trial participants.
Those communications gaps meant the medical journal was not updated, although he contended the Drugs Controller General of India was informed of the differing number of participants. “Perhaps, we could have disclosed it,” he said. “It was not intentional. It was because of these multiple teams working. … It was not so much a discrepancy. … The whole idea was, ‘Let’s get the data out.’”
“In hindsight, perhaps there should have been different communication channels,” he continued. “We were focused on the manufacturing part, safety of individuals, the ethics of handling subjects. … Our focus was in a different state of mind. … (This was) one area that skipped our attention and perhaps next time, one of the caveats is that we know we need to be careful.”
One expert suggested the differing numbers are worth noting, but may not have proven disruptive. “The discrepancy in the number of patients in Phase 1 wasn’t huge, so the question becomes would the change in patients have impacted results? Probably not,” said Reshma Ramachandran, who heads the Collaboration for Regulatory Rigor, Integrity, and Transparency at the Yale School of Medicine.
Nonetheless, the company apparently dosed a larger number of subjects than what was reported in the medical journal, explained Aisola, who suggested this raises questions about the integrity of the trial data.
The circumstances surrounding the various changes to the trial protocols are also unclear.
For instance, in September 2020 the CDSCO expert committee approved amendments to the Phase 2 protocol, including dropping the placebo arm, with two vaccine candidates being compared instead. There were also changes made to the dosing interval and the age of the participants. The stated reason was to expedite Phase 2 testing, according to the documents. The protocol was published in Lancet Infectious Diseases.
In addition, the expert committee allowed the trial to proceed on the basis of animal studies, since immunogenicity data from the Phase 1 stage of the trial had not been evaluated and, therefore, was not ready to be presented. (See minutes from meetings here and here). This type of data is important because it shows the extent to which a vaccine is producing an immune response to a virus.
In explaining these changes, Bharat Biotech’s Mohan maintained that “speed was dictating (our actions) and as data was coming along, we were getting (more) comfortable. … There was a lot of dialogue back and forth, intensive discussions. … They (the regulators) were quite tough on us. … They had similar questions: Why are you changing protocol, strategies?”
“It was extensively debated, with keeping the final objective in mind of getting a vaccine in time and not cutting any corners.”
Krishna Mohan, director, Bharat Biotech
“All of these were debated day in and day out. Whatever we did was with the clear intention of doing it right,” he continued. “There was no question of reducing sample sizes. … There were not off the cuff or random thoughts. … It was extensively debated with keeping the final objective in mind of getting a vaccine in time and not cutting any corners.”
Under the circumstances, it may be reasonable to advance the trial, according to Yale’s Ramachandran. “There’s a high rate of success with Phase 1 studies as they are usually small and meant to look at safety and dosage solely,” she said. “If there were no egregious safety events during the trial, which would have halted the trial when it was occurring, then one could assume the drug is safe before going to Phase 2.”
But she called the decision to drop the control arm “problematic.” Why? Such a move may be feasible in a situation where a vaccine is already an established product and proven to be effective. Under this scenario, it would make it unethical to have a control arm instead of a comparative group, Ramachandran explained. But that was not the case in September 2020.
“Given that Phase 2 is meant to study efficacy in addition to any side effects, a control arm is necessary to demonstrate whether due to the vaccine, the primary outcome was actually met or if it was by chance. Without a control arm for this phase, you would not have evidence of relative efficacy compared to patients not having access to anything at all,” she told us.
Moreover, such data is needed to track how population immunity changes due to rapid exposure to the virus. “Without this data, any assessment would be confounded by asymptomatic infection of the (participants),” said John. “It is unclear from the documents presented that there was a reasonable justification for either decreasing the sample size or for dropping the control arm altogether.”
Still more troubling, they said, was the decision to approve the Phase 3 protocol — including the selection of the vaccine candidate — while the Phase 2 stage of the trial was ongoing, and those results were not yet known. Phase 2 is a safety and effectiveness checkpoint for the Phase 3 participants and, therefore, is an important way to examine risks for those participants.
In other words, this is the point in the process when regulators would want to determine if a larger Phase 3 study would be beneficial, but Phase 2 data is needed to make that call. Without the data, it is difficult to design a Phase 3 study, which is used to gauge effectiveness, ensure the testing can be extrapolated to the general population, and capture long-term side effects, said Ramachandran.
Yet the minutes of an Oct. 5, 2020, meeting of the subject expert committee indicates the panel wanted the vaccine candidate to be chosen based on the safety and immunogenicity data from Phase 2, even though the data had not been presented by the company.
Another puzzle concerns details provided in protocols for the Phase 3 study.
A third version of the Phase 3 protocol, which was dated October 2020, noted the vaccine candidate was chosen based on animal studies and Phase 1 interim study data (go to page 15 for selection of formulation). However, a fourth version of the protocol — dated March 2021, which was after Phase 3 recruitment took place — claimed the candidate was chosen also based Phase 2 safety and immunogenicity data. (See page 15.) But Phase 2 immunogenicity data was not available when the third version of the Phase 3 protocol was approved and the recruitment had taken place.
“This gave the impression that the regular course of vaccine development was followed by considering data as they moved from phase to phase of the trial,” said Aisola. “They strategically updated the protocol to change the facts.”
Bharat Biotech’s Mohan, however, maintained that the company felt confident proceeding to Phase 3 as it did, given results in animal testing, and that the decision to settle on dosing was “extremely well-debated.” He denied any shortcuts were taken and that there was “a lot of pushback from regulators” to justify the company’s choices.
“There was one continuum of trial results, one continuum of decision-making and one continuum of achieving end results. As we kept building data and gaining more confidence, we seemed to be on the right track. So we didn’t wait for Phase 2. … We said, ‘Let’s start Phase 3. We can keep doing research, but we would be late.’”
“We were getting criticized for being late and there was pressure in the country. … Some of these were political and some of these were scientific. … I think the decision to move was fairly rigorous.”
Nonetheless, the discrepancies and protocol changes prompted Christian Medical College’s John to suggest that a review is warranted. “The obvious inconsistencies in the publications vis a vis their submissions is very concerning and should be looked into carefully,” he wrote us.
When asked whether the editors at Lancet Infectious Diseases will probe the published results, a spokesperson wrote to say that “we are in contact with the study’s authors and will provide further updates as necessary.”
She added that the Lancet publishing group “sets extremely high standards for publishing, and we are committed to ensuring that our editorial processes meet our standards of excellence. All original research articles published in Lancet journals undergo independent, external peer review and follow best practice guidance on publishing excellence from recommendations made by the International Committee of Medical Journal Editors and adhere to the Committee on Publication Ethics guidelines.”
Meanwhile, the episode confirms ongoing concerns about Indian government oversight of the development of medical products, according to Aisola. By greenlighting the approach taken toward approval of the vaccine, the government agencies fostered a climate in which patient safety takes a back seat to economic and political goals.
“We know the government was very much a part of the whole process. The vaccine came from a government lab and was offered only to Bharat Biotech to develop,” she told us. “Thereafter, the government was supervising and shepherding it to the point where it was rolled out with Phase 3 which wasn’t completed at the time.
“This is about cutting corners and a push to have an Indian vaccine. … We now have a deeply compromised regulatory system going forward. They did things not in the rule book and this has become the norm. And that’s the most damaging aspect about all of this.”