NOVEMBER 18, 2004
Little Rock, Arkansas
Dr. Brian Tempest and several other Ranbaxy executives huddled by the Arkansas River in plastic ponchos under a driving rain. They were determined to make the most of the soggy event, which included President George W. Bush, three former U.S. presidents, and numerous members of Congress. Ranbaxy had donated close to $250,000 so its directors could be at the opening of the William J. Clinton Presidential Library and Museum.
By their side was Dr. Agnes Varis, the founder and CEO of AgVar Chemicals, who had been serving for some time as their political chaperone in the United States. A major Democratic Party donor with a deep knowledge of America’s drug industry, she was a longtime friend of the Clintons and had taken Ranbaxy executives to parties at their Westchester home. When Clinton’s presidency ended, Varis had put a political bumper sticker on her chauffeured Bentley that read I MISS BILL. She had contributed nearly $500,000 to the Clinton Foundation for the library dedication.
Using the event as leverage, Ranbaxy did everything possible to link itself to the former president. The company released a public statement touting its “close association” with the Clinton Foundation and their shared goal of “providing pharmacotherapy to patients stricken with AIDS in financially depressed countries.” A subsequent Ranbaxy newsletter described the company as an “honored guest at this momentous event.” Though certainly an overstatement, it was true that Clinton took extraordinary steps to thank Ranbaxy for its role in making cheap HIV drugs for Africa.
Within six months of the library dedication, Clinton returned to India, where he spent more time with Malvinder Singh and other Ranbaxy executives on an AIDS panel and at a cocktail event. So much face time with the former U.S. president was like jet fuel for the company’s public image—and its bottom line.
From a distance, Ranbaxy’s ascent appeared unchecked. By early 2004, the company’s global sales had surpassed $1 billion. In the United States, Ranbaxy had become the fastest-growing foreign generics maker, with ninety-six products on pharmacy shelves and fifty more applications before the FDA. Its medicine had become important to the AIDS programs of two American presidents. And the company had big future plans: to reach $5 billion in global sales by 2012, move from the world’s eighth-largest generic drug company into the top five, and launch its own specialty products.
The company’s marketers spelled all this out in a newsletter, Ranbaxy World, which emphasized the company’s integrity and social commitment. It pointed to new quality initiatives, Ranbaxy’s elaborate code of conduct, and its dedication to making low-cost antiretroviral drugs for poor Africans. As the newsletter stated, the company’s “quest for growth and excellence goes hand in hand with unflinching commitment to integrity in all relationships with employees, customers, suppliers, government, local communities, collaborators and shareholders.”
Beneath this gauzy sentiment, glimpses of a different sort of company were visible. One week after the library dedication, President of Pharmaceuticals Malvinder Singh gave an interview to an Indian website in which he attributed Ranbaxy’s success, in part, to being a “very aggressive marketing company, fighting for a market share at rock-bottom prices and making that model work.” He didn’t explain how the model worked exactly, but went on to say, “Ranbaxy is today what it is because we took the risks.” He would later explain that Ranbaxy had risked being the first Indian pharmaceutical company to set up operations and manufacturing facilities outside of India.
But inside the company, executives were grappling with a different set of risks.
The initial disclosure of fraud at Vimta, the company Ranbaxy had hired to test its AIDS drugs, was like a teetering domino, threatening to topple interconnected drug applications approved by regulators around the world. As charitable organizations asked the company for underlying data to support its claims, the problem confronting Ranbaxy executives had become almost unsolvable. Drugs that had never been tested, or whose tests revealed a failing product, were now due to be reregistered in countries around the world. Much of the raw data didn’t match what the company had filed with regulators. Either it didn’t exist, didn’t make sense, or had been fabricated at some point. A refusal to share the data would trigger further suspicion, leaving the company with two bad options: come clean—which would have disastrous business consequences—or lie more.
The company needed to start testing drugs properly. But that not only risked exposing past fraud but often required a new set of lies. This Catch-22 played out in a torrent of confidential emails in which Dr. Tempest and future CEO Malvinder Singh were often cc’ed and also weighed in. In mid-July 2004, a UNICEF official asked why Ranbaxy had submitted only limited stability data for several of its AIDS drugs. Companies must prove that their drugs will remain stable in a range of temperature conditions. The required tests, which help to establish the shelf life of a drug and measure impurities over time, are conducted in chambers that resemble oversized refrigerators, which can replicate extremes of heat and cold.
UNICEF’s question prompted a round of panicky internal emails. In one titled “Stability Studies—Urgent,” an executive wrote, “According to UNICEF, if we fail to furnish the data by Wednesday evening and do not provide the information requested in the mail below, then we can forget about this tender.” He added, “this tender is worth $ 5 million and we cannot take nay chances on it.”
But the only data that Ranbaxy had to give UNICEF was a nonsensical hodgepodge that would raise more questions. The limited testing that Ranbaxy had done on the HIV drugs showed some impurities remaining constant or even decreasing between nine and twelve months, which was technically impossible. As an executive pointed out, these problems “will certainly raise a doubt in the mind of the reviewer . . . we need to revise this number.”
Executives grappled with similar problems in dossiers for markets around the world. In February 2005, a company executive wrote to colleagues regarding the company’s filing in Spain for the antibiotic cefuroxime axetil: “Please advice the way forward. This dossier was scheduled to go in Dec, 04. We have been waiting for your response for the last 2 months. We need to conclude this ASAP.” This email triggered a terse reply from a senior scientist: “During our discussion in Gurgaon on 27th Jan, I mentioned clearly that the data in our Archives and that of the filed one is Differing Entirely. So, I cannot send the Data.”
Months earlier, in September 2004, as the FDA evaluated Ranbaxy’s PEPFAR application for certain AIDS drugs, it asked to see data the company had filed with the World Health Organization. The director of regulatory affairs, Dr. Arun Kumar, wrote to colleagues, including Dr. Tempest: “In case we do not share the data at this stage we will be under question. . . . Reasons for not sharing the data on existing batches could be difficult to explain.” He added, “We are crossroads becoz adequate data has not been generated on the products as per WHO requirements.”
This was particularly problematic, as the company had assiduously cultivated its relationship with the FDA. In the intense debate over what data to share with the agency, Ranbaxy’s U.S. president, Dipak Chattaraj, noted in an email that the two top officials in the FDA’s Office of Generic Drugs “are so well disposed towards Ranbaxy that trying to [be] difficult with them can only cost and not help us.”
The company’s problems with the FDA were far bigger than just the drugs intended for Africa. Ranbaxy had not properly tested the stability of almost any drugs on the U.S. market. The most basic good manufacturing practices require continuous monitoring of drug quality. Drug stability must be tested at intervals called “stations”: three months, six months, nine months, and so on. So long as a drug is on the market, that data has to be filed in an annual report with the FDA. One is never out of data, because obtaining it is simply part of the process.
But the company had hit an impasse, leaving its executives to confront the fact that they had virtually no thirty-six-month stability data to submit for any U.S. commercial batches. This was more than an “oops.” It was the equivalent of trying to read a roadmap upside down—after you’ve already crashed the car into a tree.
In frantic emails, executives grappled with this seemingly insurmountable problem. Abha Pant sent a terse email to her colleagues: “This is a very serious issue. I do not know how are we going to file the Annual Reports and what reasons are we going to give to the [FDA] for not submitting the stability data. . . . We need all this data and there is no way out.”
Even as company executives resolved to start testing the drugs in “right earnest,” as one of them put it, a similar crisis was playing out in dossiers around the world. By 2005, twenty-two high-priority products faced reregistration in at least one country. All had been made at Ranbaxy’s Dewas manufacturing plant in Madhya Pradesh, and none had been tested adequately. Arun Kumar explained to his colleagues in an email, “For most of the products the data is not available, and also the archival data is not there.” In short, they had never been tested.
In assessing the task before them, a quality assurance director at Dewas wrote to his colleagues in February 2005, cc’ing Brian Tempest, “We are going to start almost on zero basis in case of majority of products where we do not have required stability data. The task seems very difficult.” In other words, drugs already on the market had to be tested from the ground up. By then, Tempest had sent the director of global quality an urgent email: “There will be no business left to pay your salaries unless we get on top of the stability work for re registrations.”
Previously, the company had taken data from fledgling research and development batches and falsely represented in filings that it came from much larger exhibit batches, which are harder to control. One troubling drug was the co-amoxiclav suspension, the antibiotic that is often used to treat ear infections in children, the same medication that had failed to cure Thakur’s son. Ranbaxy had registered the drug with a twenty-four-month shelf life in almost thirty countries, but tests showed that its shelf life was actually closer to eighteen months. In reregistering the drug, a senior consultant noted in an email, cc’ing Tempest and Malvinder Singh, it would be “useful” to have a “plausible explanation for this reduction in shelf life” when dealing with the regulatory agencies.
If most executives were seeking guidance on how best to lie to regulators, others were concerned about the fraud they were expected to commit as part of their job. Some balked at filing false data. Others flat-out refused to participate in illegal acts. However, sometimes even the most scrupulous employees ended up being drafted, unwittingly, into the company’s fraudulent schemes. Most Ranbaxy executives were expected to carry suitcases full of brand-name drugs when they traveled to India. At Ranbaxy’s New Jersey headquarters, suitcases purchased at the local Walmart were kept packed with drugs, waiting for the next traveler to India. The suitcase-toting seemed innocent enough. Most executives assumed that the drugs were needed for research and development.
Generic drug companies often study small amounts of a brand-name product in order to reverse-engineer it or to reference it as a point of comparison in applications. But proper channels for purchasing and transporting such drugs are well established and became ironclad with the 2001 passage of the Patriot Act. Personal transport of drugs was technically illegal and a form of smuggling. To the dozens of employees pressed into ferrying the drugs, often on an emergency basis, it seemed like a minor shortcut, possibly to cut shipping costs, avoid quarantine, or speed up delivery.
In one year alone, seventeen executives took undeclared drugs from the New Jersey office through Indian customs, four of them doing so multiple times. The most frequent couriers included the company’s U.S. president and even its U.S. executive director of regulatory affairs, Abha Pant, who was responsible for ensuring that the company followed the rules.
At Ranbaxy, top executives skirted these regulations and sometimes oversaw the illegal ferrying of drugs at the very moment when the company faced deadlines to resubmit data to regulators. Some executives came to suspect that the company was using the brand-name samples as a substitute for its own, in order to generate data showing how closely Ranbaxy’s drug matched the brand it was seeking to replicate. This would explain the urgency surrounding the drug runs, especially when some Ranbaxy staffers strenuously resisted being used as drug mules.
In May 2004, a regulatory project manager refused to take French brand-name samples to India. He protested in an email, “I will NOT be bring any samples with me, not only I believe is this company policy but I personally do not feel comfortable bringing samples in this manner.” An executive pushed back: “It is critical that the samples are carried by you. We cannot delay it.” The employee flatly refused.
Malvinder Singh, then the company’s worldwide head of pharmaceuticals, got involved. Through his secretary, he asked when the samples would reach Gurgaon. “These products have been sitting in our London office and it is a pity to see that nobody takes ownership of them.”
This triggered a response from the company’s president of global pharmaceutical business: “Dear Malvinder, I need to explain to you how labour laws work within Europe. As taking these samples to India is in principle illegal we cannot force people to do so. . . . Normally however we find our people willing to take the risk.” So important was this to the company’s business that the executive then went on to make an extraordinary suggestion to Singh: that since Tempest and Singh had been passing through the United Kingdom on a regular basis, “[I] would ask you to in future also make yourself available for carrying samples back.” Other senior executives were also pressed into service.
In the event that they were caught, those who carried the drugs for Ranbaxy were given a letter claiming the products were for research and development and had no commercial value. In June 2004, one executive got stopped by Indian customs with hundreds of packs of an antinausea drug, Kytril, worth thousands of dollars that he hadn’t declared. The drugs were seized. One Ranbaxy executive noted internally that “in the absence of correct documents this is considered as an illegal way of bringing the medicine in to India.”
The lies and efforts to conceal them absorbed the energies of the entire company. In August 2004, the company’s top executives, including Tempest and Malvinder Singh, held a meeting in the executive conference room. On the agenda, according to emails, was the “strategy for filling gap between requirement & availability.” In other words, how were they going to submit data they didn’t have?
In September 2004, the executives met again and opted for a solution: to move the most important manufacturing for the United States and PEPFAR from the troubled Dewas plant to the newer one in Paonta Sahib, in the hope that by severing links to the past fraudulent manufacturing, regulators would not detect it. Instead, Dewas, which had almost nonexistent quality systems, would remain the manufacturing site for the least-regulated markets, including Brazil, Mexico, Vietnam, and elsewhere.
Publicly, company executives spun this change as a response to big demand from the American market and PEPFAR. In early January 2005, at a meeting of Indian generic drug makers and AIDS activists in Mumbai, Ranbaxy’s HIV and essential drugs project manager, Sandeep Juneja, explained to the meeting attendees that the company’s new strategy would result in its drugs being quickly approved by PEPFAR and reinstated on the WHO list. “We wanted to harmonise everything to the U.S. market and consolidate manufacturing in one place,” he said. “If a product is approved by the FDA, then it is accepted anywhere.”
Two days later, Arun Kumar wrote to a UNICEF official, explaining the shift in the manufacturing site for lamivudine, an AIDS drug: “We have changed the site of manufacture of the product from Dewas to Paonta Sahib facility to facilitate handling high business requirements.” Four days after making that claim, however, as the company prepared to resubmit the data for its antiretroviral drugs to the WHO, Juneja reiterated the company’s real strategy in an email, cc’ing Tempest. “We have been reasonably successful in keeping WHO from looking closely at the stability data in the past,” he wrote, adding, “The last thing we want is to have another inspection at Dewas until we fix all the process and validation issues once and for all.”
But the shift to Paonta Sahib was far from a perfect solution. In some cases, the new studies weren’t going to be completed in time to meet new registration dates. As one executive asked in an email, how were they going to handle the “interim situation,” where regulators were demanding new data that wasn’t ready yet? If the data came from Dewas, “how do we use them in the Paonta dossier?” The answer from a colleague came back: “pl speak to me over the phone. I shall be able to answer yr query.” The answer was almost certainly interim fraud, or temporarily representing old data from Dewas as new data from Paonta Sahib.
However, as Ranbaxy geared up to start testing drugs at Paonta Sahib and redirect the world’s most important regulators to that plant, the company wanted to leave nothing to chance. And legitimate testing—where drug batches can fail and formulas can become unstable—was the ultimate game of chance. Investigating the why is difficult and costly. Good manufacturing practices are so laborious precisely because they require innumerable steps to make an uncertain process more certain. If you have to actually test drugs, how can you best control the outcome? How do you put your finger on the scale each time in order to generate perfect data?
Ranbaxy came up with a rather ingenious, well-hidden solution. Whether it remained hidden would depend, in large part, on the next FDA investigator who walked into the plant. Would it be someone content to look merely at the well-polished surface? Or someone committed to piecing together a mosaic of clues? The company had no way to control which investigator showed up.
But two months after Raj Kumar made his ill-fated presentation to Ranbaxy’s board of directors, the company had a stroke of luck. The next FDA investigator to arrive at Ranbaxy’s Paonta Sahib manufacturing plant in Himachal Pradesh was Dr. Muralidhara B. Gavini.
At the FDA, Muralidhara Gavini, who was widely known as Mike, had distinguished himself in one important respect: he was one of the few investigators not only willing but happy to do foreign inspections in India.
Most of the agency’s investigators did not want to travel there, especially as their colleagues returned with stories of broiling heat, incessant rain, harrowing traffic, hours spent traveling to distant manufacturing plants over pitted, washed-out roads, and the ubiquitous threat of getting sick from unclean water or contaminated food, which forced them to bring suitcases of peanut butter and granola bars. The difficult travel had contributed to a growing crisis at the FDA.
In theory, the agency tried to inspect every facility making drug ingredients for the U.S. market roughly every two years, whether the plant was in Maryland or Mumbai. But the FDA’s actual rate of inspections overseas was closer to once a decade, and the backlog of applications from foreign drug facilities was growing rapidly. The FDA had no obvious solution for who to send to inspect overseas plants and how to pay for it all. The agency was so desperate that it had even explored doing remote inspections, in which the plants would supply videotapes of their facilities. The proposal for inspections by videotape cited “ever-decreasing resources.”
In a system starved for volunteers, Mike Gavini was in high demand. The area around Hyderabad was beginning to develop, and his inspections took him back to his old stomping grounds. He’d grown up in Guntur, south of Hyderabad. Sending Gavini there did not violate any agency rules, in part because there weren’t too many at that point. But it did violate a basic principle. As one of Gavini’s colleagues would later observe, “You never send people to their native area. That’s just Regulatory 101.” Such an arrangement could invite corruption and compromise. But Gavini was eager for a homecoming.
He had first left India in 1972 to pursue a PhD in chemistry at the University of Arkansas. For him, America and the town of Fayetteville, were alien environments. But his work was good enough to land him a job at the Woods Hole Oceanographic Institution in Massachusetts, where he continued his research, tracing plutonium isotopes in rainwater. At Woods Hole, he uncovered traces of curium in the sediment of Lake Ontario and presented his findings at a national meeting of the American Chemical Society. Unwittingly, he had revealed runoff from a secret nuclear facility. Amid recriminations, his career as a scientist ended.
With three children to support, he spent the next ten years at a laboratory testing company until a management change left him looking for work again. In 1996, he took a 70 percent pay cut and joined the FDA’s New Jersey district office, where he got paid $35,000 a year to visit dairy farms and look for evidence of bovine spongiform encephalopathy (mad cow disease). With his PhD and a decade of private-sector experience, he was an “oddball,” as he recalled. By 1999, he was promoted to a position as a compliance officer at the FDA’s Center for Drug Evaluation and Research (CDER).
He may have been out of his depth in Fayetteville and given short shrift at the FDA, but in India he was a bigwig. Not only did he have a PhD, which automatically commanded respect, but he also represented the world’s most powerful regulator. Most of the time, he went alone to the plants—an inspectional “one-man army,” as he put it. He—and he alone—could determine whether to clear a plant to export its drugs to the United States, a project he undertook without clearly defined rules.
Domestic inspections had long followed a clear formula. The FDA’s investigators would show up unannounced at U.S. factories and stay as long as they needed to follow data trails wherever they led. “We walk in, show the badge, give them notice of inspection,” as FDA investigator Jose Hernandez put it. The relationship with the facility was also clear: there was none. “Those guys [in New Jersey] won’t even take a cup of coffee,” as one investigator put it.
But the rules for overseas inspections were muddy at best. Seeking to avoid confrontations that might involve a foreign government and lead to an international incident, the FDA prioritized diplomacy over confrontation. It announced its visits to the plants weeks, even months, in advance, and relied on the companies to act as hosts and travel agents for its investigators, booking hotels and ground transportation. This suited Gavini.
The approach he chose was collaborative. He viewed himself as a partner, educating the companies as to what constituted robust quality systems. “He conducted himself more as a consultant,” one of his colleagues noted. Gavini regarded the FDA and the companies as sharing the same objectives. “We’re not on opposite sides of the table,” he said. “Quality is our goal.”
He tried to improve the companies’ techniques. He taught them how to clean manufacturing equipment properly and stressed “common sense.” The industry, he said, “learned a hell of a lot from me.” His personal measure for each inspection was, “Have I contributed positively at this company? The net result should be positive.” It certainly was. As he inspected, the pharmaceutical sector in Hyderabad grew under his watch, churning out active ingredients and ultimately becoming the bulk-drug capital of India.
Gavini put in ten- to twelve-hour days. He expressed disdain for colleagues who worked shorter hours or filed their reports months after the fact. But on the rare occasion when colleagues accompanied him, they were dismayed by what they saw: the chumminess of his relations with the companies he was inspecting; his direct communications with the executives, in person and by phone; and his tendency to let companies off the hook. He would even send draft versions of his regulatory findings, called 483 reports, to the companies for review before he submitted them officially.
Gavini rejected what he saw as needless secrecy. “I don’t know why FDA investigators keep everything to themselves,” he said. “I discuss every aspect I am writing [with the companies].” This gave companies the opportunity to influence his findings. If the firm said that it would fix its problems, that was good enough for Gavini. In one 2003 inspection report, he even documented how a plant’s managing director had pledged improvements by phone, a promise that he deemed “satisfactory.”
The honor system was a bad way to improve drug quality—at least, according to a report by the U.S. Government Accountability Office (GAO) written in 1998, two years before Gavini started his inspections. The report bashed the FDA for downgrading inspectional findings based on “foreign manufacturers’ promises” to make changes. “As a result, FDA conducted fewer reinspections of these facilities to verify that foreign manufacturers had corrected serious manufacturing deficiencies.”
Gavini was not inclined to treat his countrymen like criminals, he would later explain. But as the pharmaceutical sector in Hyderabad exploded—and FDA investigators returned to facilities that Gavini had previously cleared to find grievous violations—his reputation as a lax investigator spread. Among his colleagues, he came to be known as the quintessential “NAI Inspector”: one whose most common finding was No Action Indicated.
On December 17, 2004, to the delight of Ranbaxy’s executives, Mike Gavini, an investigator they knew well, arrived at the Paonta Sahib plant. He was there for a preapproval inspection to make sure the facility could adequately make two AIDS drugs, lamivudine and zidovudine, for the U.S. AIDS relief program in Africa.
His arrival was not a surprise to the company. As was typical for the FDA’s foreign inspections, a visit had been announced weeks in advance, and the company was involved in planning it. In his summary of the inspection, Gavini noted: “The firm provided transportation to and from the plant” and “accommodations were provided.”
Gavini stayed for five days. Trailing him throughout the plant were many of the same executives who’d been sending panicked messages to one another about the lack of data. Gavini saw none of that. In his inspection report, he noted some unclear instructions in batch records and unclear cleaning procedures. However, he gave unqualified praise to the company’s program of stability testing, which exposes drugs to different temperatures and humidity levels to see how fast they degrade in different environments and what their expiration date should be. He noted: “Stability sample traffic in and out of the chambers was monitored and the stability inventory maintained properly.”
To come to this conclusion, Gavini must have walked directly past a Thermolab stability refrigerator that had been installed seven months earlier. The large walk-in refrigerator was maintained at 4 degrees Celsius. Use of the refrigerator was not noted on any of the company’s applications with the FDA. Had Gavini opened it, he would have found hundreds of bottles of samples that had not been recorded, stuffed inside cardboard boxes.
Why Ranbaxy was using the refrigerator would become one of the most contentious, bitterly fought questions for years to come. But again, Gavini didn’t raise any questions. To Ranbaxy’s nervous executives, he was a dream investigator, either unable to find or not on the lookout for illicit solutions to drug testing.
Not surprisingly, his findings were minimal. He gave the plant a clean bill of health and determined that it was in “overall compliance” with current good manufacturing practices. His conclusion was NAI: No Action Indicated.