The dangerous failure to stop tainted remdesivir

This was the second such incident in a week. Just days ago, Pal had seen the same happen to four patients at Jhansi’s Lifeline Hospital, where he also consults. And there was a common thread; the patients at both hospitals had received remdesivir from the same batch, coded V100167. It was manufactured by the Gujarat-based firm Zydus Cadila.

Pal and his colleagues realized that some ingredient in this batch of the popular antiviral was triggering the reaction. But it was tough to say what. It was a chaotic period, Pal recalls. Hospital beds were full, and doctors had little time to investigate further. “Managing so many covid patients was already a challenge. So, when the drug-reaction occurred, all we could do was to report it and treat it.”

Assuming that the medicine had somehow triggered the patients’ immune system, the doctors prescribed an antihistamine and a corticosteroid, both drugs that counter drug allergies. They also stopped using Cadila’s remdesivir, which had been procured by the Uttar Pradesh government. Instead, they told families of other covid patients to purchase a different brand. Once the drug was discontinued, patients seemed to improve, the anesthesiologist told Mint.

But what happened in Jhansi was just the tip of the iceberg. Unknown to Pal, in May 2021, over a dozen hospitals across Uttar Pradesh, Rajasthan, Maharashtra, Gujarat and Bihar reported patients falling similarly sick after getting remdesivir. The batches and formulations overlapped, and the manufacturer was always Cadila.

Not everyone recovered from the seemingly-tainted drug. In 69 reports of adverse events from Uttar Pradesh that Mint reviewed, doctors recorded the death of one patient. This number is likely an underestimate, because the doctors filing these reports didn’t always note whether the patient recovered fully from the symptoms.

Despite the obvious link between Cadila’s batches and the life-threatening adverse events, few state drug regulators responded as they were supposed to. Mint’s investigation found that regulators acted in a haphazard manner, declaring the batches to be of standard quality without testing them fully, and without seeking an explanation from the company.

Only the Bihar regulator identified the cause of the mysterious illness. Their testing found Cadila’s batch V100167 to contain bacterial endotoxins—compounds present in the protective envelopes of bacteria, which cause fever, chills and life-threatening septic shock in humans. Regulators in the remaining four of the five affected states never tested the medicines for endotoxins, although low endotoxin levels are a key quality requirement for drugs that are injected into humans.

The Gujarat regulator, the Gujarat Food and Drug Control Administration (FDCA), which bore the greatest responsibility in the episode, was missing in action altogether. Cadila’s manufacturing units are located in Gujarat, which means that the FDCA was obliged to follow up on adverse-event reports by testing the batches, asking Cadila to investigate the cause of the illnesses, and ensuring the batches were taken back from customers across India.

Yet, Mint has learnt that, despite being informed about the issue, the FDCA’s laboratory never tested the implicated batches. Nor is there any evidence that, as required under Indian law, the FDCA or Cadila recalled the tainted drugs.

The Gujarat FDCA did not respond to emails from Mint.

A spokesperson from Cadila denied all reports of adverse events, and did not answer questions about recalls.

Same batches, same symptoms

It was during a small window in the first two weeks of May that multiple hospitals across Uttar Pradesh began seeing patients deteriorate with remdesivir. Apart from the Chiranjeev and Lifeline Hospitals in Jhansi, these included Jhansi’s Maharani Laxmibai Medical College, Meerut’s Lala Lajpat Rai Memorial Medical College, Moradabad’s Teerthankar Mahaveer University, Lucknow’s Sanjay Gandhi Postgraduate Institute of Medical Sciences and Varanasi’s Ashirvad and Apex hospitals.

Anuj Kumar, a dermatologist at Teerthankar Mahaveer University’s hospital, was deputed to a covid ward then. Given the shortage of medical staff, everyone was on covid duty, regardless of their specialization.

Between 7 and 9 May, Kumar personally attended to five patients who received batch V100156 from Cadila, before turning feverish and breathless. Kumar says it was easy to tell that the medicine was at fault, even though covid causes similar symptoms. “Say, someone is stable on five litres of oxygen, and then their SPO2 drops within an hour of getting remdesivir, it’s clear what is happening. And it isn’t like this happened to only 1-2 patients. It happened to quite a few.”

In the adverse-event reports he filled out, Kumar listed four patients as recovered, but didn’t record the same about a 39-year-old male, who broke into shivers after the injection. Kumar says that he and his colleagues didn’t always know how their patients eventually fared, especially when they were transferred from the ward due to deteriorating health. “If SPO2 dropped below 70% or so (after remdesivir), we would admit them to the intensive care unit. After that, the patient could be shifted to other wards, and we didn’t really have an idea of what happened to them,” he told Mint.

Also Read: What are the quality specifications that drug makers need to follow

In total, Mint accessed 15 adverse events reports associated with V100167 and V100156 from Teerthankar Mahaveer University. Out of the 15 cases, doctors did not list the outcome for five patients, while they recorded one death, and nine recoveries. Asked if the unrecovered patients were followed up, Ajay Pant, the medical superintendent of the hospital, said no records of the event were with the hospital today, because the institution had been taken over by state government officials in May, who subsequently took all records away with them.

This story repeated itself across India in May. In Rajasthan, the Rajasthan University for Health Sciences (RUHS) and the Employees’ State Insurance Corporation (ESIC) Model Hospital, both in Jaipur, sounded the alarm about batch L100148, a powdered version of remdesivir, to be mixed with sterile water and saline before use. The same batch had triggered reactions at the two Varanasi hospitals, and at Aurangabad’s Government Medical College too. In Bihar, multiple Patna hospitals complained about V100167. In Gujarat, hospitals run by the Ahmedabad Municipal Corporation, as well as private hospitals, red-flagged V100170.

In the majority of cases, government-owned procurement companies had purchased the medicines. These included the Rajasthan Medical Services Corporation Ltd, Haffkine Biopharmaceutical Corporation Ltd, Uttar Pradesh Medical Supplies Corporation Ltd, Bihar Medical Services & Infrastructure Corporation Ltd and the Gujarat Medical Services Corporation Ltd. However, a few private hospitals had purchased the drugs directly from Cadila’s distributors too.

Most hospital-staff Mint interviewed claimed that the incident didn’t result in deaths. But, as in the case of Teerthankar Mahaveer University, this claim likely reflects the fact that patient outcomes were poorly tracked. Even when patients did die after receiving the drug, in the confusion of the pandemic, it was hard to attribute the death to either underlying covid co-morbidities or to the medicine, a doctor from an Ahmedabad private hospital said.

Four states drop the ball

The strong association of the symptoms with specific batches prompted several hospitals to inform their respective state drug regulators. This is when the regulators dropped the ball.

In response to the reports, the regulators ought to have conducted the entire battery of quality tests for medicines (See box: What are quality specifications?) to understand what was driving the illnesses. And because patient symptoms matched so closely with those triggered by endotoxins, two quality tests were key: a so-called sterility test and an endotoxin test. (See box on the next page). Instead, Mint found that two state regulators did not undertake any quality testing at all, while one skipped the endotoxin test, and another skipped both endotoxin and sterility tests.

The reasons why four states failed to get to the bottom of the drug-toxicity differed from place to place. Sometimes, it had to to do with the state’s defective adverse-event reporting system and the negligence of state’s drug-procurement agency, while at other times, it had to do with an ill-equipped drug lab, confusing rules under the Drugs and Cosmetics Act, and an outdated Indian Pharmacopoeia—the book that sets minimum quality specifications for medicines sold in the country.

For instance, in Rajasthan, the adverse event reporting system was so dysfunctional that news of the patients falling sick from batch L100148 at the RUHS and ESIC Model hospitals never reached the state drug regulator. While the doctors at these hospitals claim they informed the Rajasthan Medical Services Corporation, the corporation didn’t pass this information onto the regulator, which is tasked with evaluating drug quality. “No such case came to our notice,” Raja Ram Sharma, the head of Rajasthan’s drug regulatory agency, told Mint.

Instead, according to Alok Ranjan, the former managing director of the corporation, (he was transferred to the state’s energy department), his team merely contacted Cadila for an explanation in May 2021. Cadila responded saying they had not seen similar adverse events elsewhere—an incorrect statement, given that two Varanasi hospitals and one Aurangabad hospital had reported the same issue with L100148 at roughly the same time.

Further, Cadila allegedly suggested that the saline mixed with the remdesivir powder may be to blame for the symptoms. “Cadila wrote to us saying: you might also get your saline tested, because it might be due to the saline water you are using,” Ranjan said.

His team accepted Cadila’s explanation without any independent testing. However, Cadila did replace all unused bottles from L100148 “immediately”, at no extra cost, said Ranjan. The Rajasthan Medical Services Corporation’s current chairman Vaibhav Galriya did not answer Mint’s queries.

A similar tale of negligence played out in Uttar Pradesh, too. Unlike in Rajasthan, multiple hospitals here did inform the regulator directly. In turn, the regulator decided to test batches V100153, V100156, V100166 and L100148 in its Lucknow lab. But the lab lacked facilities for either the sterility or the endotoxin tests, according to an official at the lab who didn’t want to be identified. The net result: the lab declared all the batches to be of standard quality, without even conducting what were arguably the most crucial analyses.

What happened in Gujarat, the state where Cadila’s manufacturing plants are located, is a mystery. The state drug regulator was informed about adverse reactions not just by local hospitals, but likely also Cadila itself. Indian law requires every pharmaceutical manufacturer to inform the local regulator about all adverse-events reports it receives, and Mint has independently confirmed that Cadila was sent such reports by Uttar Pradesh and Bihar.

Yet, the Gujarat regulator ostensibly never collected any of Cadila’s implicated samples for testing. Joint commissioner of Gujarat’s regulatory labs, Jyotsana Makwana, told Mint that no samples from V100153, V100156, V100166, V100167, V100170, or L100148 ever arrived at the lab. Repeated questions to the commissioner of Gujarat FDCA, Hemant G Koshia, on why his department failed to test samples, went unanswered.

M’rashtra: law falls short

The Maharashtra regulator’s story is somewhat unique. It received adverse events reports from hospitals, had the facilities to test the remdesivir fully, and ostensibly, even the intention to do so. Yet, it failed to conduct an endotoxin test, mainly because it was following outdated sections of both the Drugs and Cosmetics Act and the Indian Pharmacopoeia.

To understand exactly what happened in Maharashtra’s case, one needs to understand the concept of a “patent or proprietary drug”. Frequently, the quality specifications for important drugs sold in the country are listed in a text called the Indian Pharmacopoeia. The Drugs and Cosmetics Act makes it mandatory for manufacturers to comply with these specifications (see box).

But many drugs have no entry in the Indian Pharmacopoeia, either because they have just been introduced into the country, or because they are not widely used, or because the Indian Pharmacopoeia has simply not gotten around to creating such specifications. Such drugs, for which no common, official quality specifications exist, are called “patent or proprietary” drugs.

The powdered remdesivir sold by Cadila in May 2021 was a patent or proprietary drug, because the Pharmacopoeia was yet to impose specifications for it. And in this situation, the Drugs and Cosmetics Act has a giant loophole. For such drugs, the act merely requires that the drug comply with the Pharmacopoeia’s general quality specifications, also known as “general monographs” for the dosage form of the drug.

In other words, if remdesivir is sold as a powder, it must comply with the Pharmacopoeia’s general monograph for powder for injection. And this monograph had no test for endotoxins.

Several experts told Mint that this situation is ripe for fiascoes such as the remdesivir one. General monographs for drugs, capsules and injections are often not comprehensive. In the case of powders-for-injection, the Indian Pharmacopoeia’s general monograph was particularly outdated. The lack of an endotoxin test is “not scientific”, Ganadhish Kamat, who retired as the global head of quality at Dr Reddy’s Laboratories, said.

Given the shortcomings of the Indian Pharmacopoeia, when India’s apex drug regulator, the Central Drugs Standard Control Organisation, approves a drug for use in the country, it often requires the manufacturer to adhere to a more stringent list of specifications. Considering all this, government labs ought to seek specifications from the manufacturer, instead of using the general monograph, added Kamat.

The fact that manufacturer specifications tend to be more comprehensive than the Indian Pharmacopoeia’s general monographs is implicitly recognised by a few government drug labs. For instance, officials from the Gujarat drug lab told Mint that they always sought specifications from manufacturers. But no part of the Drugs and Cosmetics Act explicitly requires this. As a result, other drug labs, including those in Karnataka and Kerala, continue to rely on general monographs for patent or proprietary drugs.

This is what the Maharashtra drug lab in Mumbai did too, when it tested batch L100148 in July 2021. Following the general monograph, it skipped the test for endotoxins, Maharashtra FDA joint commissioner DR Gahane said. And because the sample complied with the tests listed in the general monograph, the lab declared the remdesivir to be of standard quality, and dropped the issue.

Bihar pinpoints the problem

Eventually, the only state regulator that uncovered the reason behind the countrywide trail of sickness was that of Bihar. As luck would have it, the systems that failed in other states did work here.

Local hospitals informed the regulator about the adverse effects, and the regulator collected V100167 for testing. Further, recognizing that the state’s drug lab did not have the facilities to conduct sterility or endotoxin tests (just like the UP lab), the regulator sent the samples to a central government lab in neighbouring Kolkata, the Central Drugs Laboratory.

What’s more, the Central Drugs Lab asked Cadila for its specifications, instead of relying on the Indian Pharmacopoeia’s general monograph. Cadila, which was manufacturing the drug under a voluntary license from United States’ company Gilead Sciences, was following more up-to-date specifications than the general monograph listed. In other words, Cadila’s specifications included the much-needed endotoxin test (although the contamination of Cadila’s batches suggests the company may not have been implementing these specifications).

Upon testing the remdesivir, the lab found that batch V100167 was failing the endotoxin test—a viable explanation for the symptoms patients had been experiencing. This led to the Bihar regulator informing the Gujarat regulator, and filing a case against the firm in a Patna Civil court in October. So far, the court has held one hearing in this case, with the next scheduled in January 2022.

Worryingly, however, there is no clarity about whether the Gujarat regulator followed up. Once the Gujarat regulator was informed, the Drugs and Cosmetics Act required that either Cadila or the regulator recall every remaining bottle of V100167 from consumers. Yet, in its email to Mint, Cadila denied all contamination, while the commissioner of Gujarat FDCA, Hemant Koshia, did not respond to queries on the issue.

Troubling questions

The handling of the situation by various states raises many troubling questions about how well India is equipped to ensure injection safety.

Perhaps, the most disquieting is the fact that both Bihar’s and Uttar Pradesh’s regulatory labs, which cater to two of the most populous states in the country, have no ability to conduct either sterility or endotoxin tests. This means the state may have been missing repeated instances of life-threatening drug contamination for years. While central labs can pick up some of the slack for ill-equipped state labs, they cannot bear the entire burden of testing in such large states.

The remdesivir incident also highlights the negligence of government procurement companies in ensuring the quality of the medicines they purchase from pharma companies.

Out of three state procurement agencies who answered Mint’s questions, officials from the Gujarat Medical Services Corporation and Maharashtra’s Haffkine Biopharmaceuticals Ltd never even learnt about the quality issues with their drugs. The hospitals they supplied the remdesivir to simply didn’t tell them that they had run into possibly contaminated batches. This raises questions about whether these agencies had any internal adverse-event reporting system in place.

Meanwhile, the Uttar Pradesh Medical Supplies Corporation and the Rajasthan Medical Services Corporation did learn about the adverse events from hospitals, but still didn’t inform the local drug regulators. Instead, they left the job to hospitals, which hospitals sometimes neglected to do, as in Rajasthan’s case. As a result, the substandard remdesivir kept flying under the radar there.

This laxity of government medical procurers when it comes to quality issues is a known problem, and is reflected in the high proportion of substandard drugs seen historically in government supplies. A 2014 countrywide survey by the National Institute of Biologicals found that around 10% of drugs from government agencies, such as the Central Government Health Scheme and the ESIC, were substandard, compared to around 3% sold in retail outlets.

An outdated Drugs and Cosmetics Act and Indian Pharmacopoeia also played a role in the remdesivir fiasco. Why did the Drugs and Cosmetics Act not require government analysts to seek up-to-date quality specifications from manufacturers? And why does the Indian Pharmacopoeia’s general monograph for powdered injections not have an endotoxin test? This gap in both the law and the Pharmacopoeia means that, in May, multiple drug labs were unable to detect safety issues with remdesivir. Rajeev Raghuvanshi, scientific director of the Indian Pharmacopoeia Commission, the body that writes the Pharmacopoeia, agreed that this was a shortcoming in the text, but said it was currently being rectified.

Finally, the Gujarat drug regulator’s central role in the entire mess cannot be understated. As the regulator which licensed Cadila’s manufacturing units to make remdeisvir, it was Gujarat’s job to ensure that Cadila met all quality specifications, and to ensure it did not sell substandard drugs in the first place. Once evidence of substandard drugs reached the regulator, it was required to investigate thoroughly, and recall all contaminated batches, given the possibly life-threatening nature of the adverse events. Instead, the state didn’t even test the batches, and it’s unclear if it recalled them.

These substantial failings of multiple agencies came together in May 2021 to make hundreds of already sick covid patients even sicker. The irony of the incident is that, as early as October 2020, results from the WHO’s Solidarity trial showed remdesivir to be ineffective against severe covid or death. The findings, however, did not influence the use of remdesivir in India, with doctors continuing to prescribe it widely.

This means that patients who received remdesivir in May likely didn’t need the drug in the first place. But many of them got it anyway, and a contaminated version to boot.

Priyanka Pulla is a Bengaluru-based reporter covering health. Her reporting on covid-19 is supported by a grant from the Thakur Family Foundation. The Foundation exerts no editorial influence on her work.