MAY 25, 2007
Cleveland, Ohio
Dr. Harry Lever, a sixty-two-year-old cardiologist, always listened to National Public Radio as he drove his half-hour commute into work at Cleveland Clinic. His drive took him down suburban back roads into Cleveland’s early-morning traffic. But this morning he was so engrossed in a radio segment that he barely noticed the streets going by.
The story described how the United States was importing huge amounts of adulterated food and ingredients from China that the FDA was struggling to inspect. The list was harrowing: toothpaste with ingredients found in antifreeze; fish bred in polluted waters and fed banned veterinary drugs; herbal tea leaves dried with exhaust fumes from trucks using leaded gasoline. The FDA inspected only around 1 percent of all the food and ingredients that entered the United States, the story explained. Imports from China were by far the most likely to be seized as unfit for human consumption. A former FDA deputy commissioner, William Hubbard, explained that investigators often blocked products that looked rotten or smelled of decay. “Filthy” was the official term. The FDA was so understaffed, however, that “only a fraction” of “filthy” food was caught and stopped at the border, said Hubbard. The rest was “slipping through the net.”
Lever parked in the hospital garage with the radio still on. His pager beeped, but still he sat there. Until that morning, he’d had no idea that so many products—apple juice, garlic powder, honey, hot dog casings, vitamin C—came into the country with such perilously weak regulation. Pets had died from tainted food. To Lever, it sounded evil.
He got out of the car but couldn’t get the radio broadcast out of his head. That night he started doing research in his own pantry, where he found a bottle of garlic powder made in China. Then he noticed that it had been certified as kosher by the Union of Orthodox Rabbis. His first thought was, a counterfeit certification. He went to the Union’s website. Sure enough, rabbis were certifying Chinese imports. But how could they know what was safe and what wasn’t? He called the Union and got a rabbi on the phone who confirmed the certification. Lever even tracked down William Hubbard, the former FDA commissioner who had been on the radio program. Hubbard verified the details, down to the tea leaves coated in lead exhaust.
Lever’s nature was “almost too passionate,” as his cousin, who was also a patient of his, commented. “He just gets furious when he knows something is not being done right.” But Lever didn’t dwell for too long on the problems of “filthy” food. He’d soon turn his attention to adulterated medicine.
Not long after hearing the NPR radio segment, Lever noticed that some of his patients were suffering from a low platelet count after taking heparin. He began raising concerns with colleagues. When it emerged that the heparin had been contaminated in China, one doctor called him a “prophet.” In his mind, he was just connecting the dots. As he did, his concerns grew. He realized that he could no longer assume that the drugs his patients were taking would work as intended.
Lever specialized in treating hypertrophic cardiomyopathy (HCM), a condition in which the heart’s muscle tissue thickens and potentially restricts blood flow. The disease can strike without warning and is the leading cause of sudden cardiac death among young athletes. Over the years, Lever developed one of the nation’s biggest practices in treating the disease and helped uncover new methods to identify it early. At Cleveland Clinic, he worked from a small ground-floor office, where mementos from grateful patients lined the walls. Medical records lay in heaps on his desk, and Post-It notes were stuck to the cabinets. Most days he took lunch at his desk, eating the salads his wife had prepared, as he reviewed his patients’ echocardiograms.
Lever liked to visualize each patient’s case as an algebraic equation. A new symptom put an unknown variable, an “X,” into the equation. Solving for X could be relatively straightforward if the other variables were known. The brand-name drugs that Lever had prescribed for years were known variables: they had almost always worked as expected. Lever’s best defense for his patients was medication: beta blockers and calcium channel blockers that could offset his patients’ irregular heartbeats or help lower their blood pressure. Many of his patients also needed diuretics, pills that reduced swelling and fluid buildup. After the radio segment, Lever started to realize that some patients he had stabilized on those drugs developed symptoms again when they were switched to certain generics. Those generics seemed to be a new X that threw off the whole equation.
He started running Google searches. If a drug roused his suspicion, he researched the drug maker and the locations of its plants—basic information absent from the drug’s packaging or product label. With increasing frequency, he also contacted Cleveland Clinic’s high-level pharmacists, who routinely gathered data from manufacturers and the FDA as part of their research into which drugs the hospital should use. He developed a sense for which drugs, or which companies, to avoid. Ranbaxy was among them.
One night his cousin, who suffered from heart disease, called him and complained that he was feeling terrible. One of the medications he took was generic furosemide, a diuretic that helped him shed excess fluid. Lever immediately asked, “What are you taking?” His cousin read the label on the pill bottle: Ranbaxy. Lever switched his cousin to what he considered a better generic made by Teva, an Israeli drug company. His cousin lost about fifteen pounds of fluid in a week.
Another of his patients, Martin Friedman, a theater professor, was having trouble keeping off excess fluid, despite taking a diuretic. His ankles were swollen, and the only way he could sleep was in an upright position, propped on pillows. Lever soon identified that his diuretic, torsemide, was manufactured by the Croatian company Pliva. Lever switched Friedman to Demadex, the brand version originally manufactured by Roche (and later purchased by the European firm Meda). Again, Friedman immediately lost excess fluid. “It was pretty weird,” Friedman noted. “It started working right away.”
The list of sick patients and suspect drugs continued to grow. One patient, Karen Wilmering, had obstructive HCM, which blocked the blood flow out of her heart’s left ventricle. For years, to control her cholesterol, she’d taken brand-name Pravachol, made by Bristol-Myers Squibb, until she was switched to a generic, pravastatin sodium. When Lever ran another test, Wilmering’s cholesterol results were alarmingly high.
Lever asked Wilmering which company made her cholesterol drug. When she told him it was Glenmark, Lever “almost jumped out of the chair, because Glenmark is from India,” Wilmering recalled. Lever dialed Wilmering’s pharmacist and insisted that she be switched to Teva’s version, which he believed would work better. In less than a month, her cholesterol dropped back to the normal range. Even Lever was stunned by the new numbers. “His jaw dropped,” Wilmering said. The following year, Glenmark recalled thousands of bottles of its pravastatin sodium, due to patient complaints that the medicine gave off a strong fishy odor. A Glenmark spokesperson said the recall was “voluntary, proactively initiated by Glenmark, and not related to product efficacy.”
Some of Lever’s patients became as outraged as he was. Christine Jones, fifty-four, who had previously run a consumer relations division at PepsiCo, retired early owing to her HCM. She was put on the beta blocker Coreg, made by GlaxoSmithKline. But after she was switched to the generic version, carvedilol, which cost $10.87 for a four-month supply, her health deteriorated. “I was having way more shortness of breath and irregular heartbeats that were waking me up at night,” she recalled. “This went on for months.” She attributed the problems to her diet and lack of sleep.
Lever immediately fingered the generic. It was made by the Indian company Zydus. “If he hadn’t put two and two together,” said Jones, “it would never have entered my mind.” Lever switched her back to the GSK brand, which cost $428 for a six-month supply. She felt better almost immediately after taking it. She began to research Zydus and found a trail of aggrieved patients posting to online forums. Alarmed by these complaints, she called the company “multiple times” to ask if its generic was just as effective as the brand, but could never get what she felt was a clear answer. The fact that the Zydus drug worked so poorly, and yet the FDA had approved it, seemed “unconscionable” to Jones.
As Lever made life-altering adjustments—switching patients from bad generics to better ones or back to the brands—he was no longer just diagnosing his patients. He was following the trail of the global economy and trying to diagnose the drug supply. The preponderance of problematic drugs seemed to be manufactured in India, but some were American-made. Many of the ingredients used in the finished drugs came from China. Lever was discovering that a complex supply chain made it extremely difficult to know where drugs had been made, who made them, and which ones worked best.
With some drugs, the problems were stark. When his patients were switched to certain generic versions of metoprolol succinate, a beta blocker, Lever noticed that they frequently developed chest pain and their heart rate and blood pressure became harder to control. The brand version, Toprol XL, made by AstraZeneca, had a time-release feature that helped the active ingredient stay in the blood, but that was under a separate patent from the drug itself. In 2006, as generic companies started selling versions of Toprol XL, they had to engineer their own mechanism for releasing the drug into the bloodstream.
That year, Sandoz, a subsidiary of Novartis, in Switzerland, started marketing the first generic version. The following year, Ethex, a subsidiary of KV Pharmaceutical in St. Louis, Missouri, also began selling a version. At the time, only Par Pharmaceutical, based in New York, had been authorized by AstraZeneca to sell a generic version, so it used the same time-release mechanism as the brand. Lever became certain that there was something wrong with the Ethex and Sandoz versions. “You get this feeling like you smell a rat,” said Lever. Sure enough, in March 2008, the FDA inspected one of Sandoz’s plants and found “significant deviations.” By November 2008, after a stark warning letter from the FDA, Sandoz quietly recalled its drug.
Around the same time, a vigilant Cleveland Clinic pharmacist sent the package insert for Ethex’s metoprolol succinate to Lever, with one phrase out of the nine-page document highlighted: “Does not comply with the dissolution test of the USP monograph.” Lever was amazed. If the drug did not dissolve according to the agreed-upon standards set by the USP, why had the FDA approved it? Together, he and the pharmacist called the FDA’s Office of Generic Drugs from Lever’s office. They wanted to know: Were patients getting too much of the drug? Not enough? They couldn’t get a straight answer. Then, on January 28, 2009, Ethex announced a sweeping recall of more than sixty products, and metoprolol succinate was among them. Ethex later pleaded guilty to two felonies and agreed to pay a fine of over $27 million.
Lever had no official data to support his observations. He couldn’t see inside the drug plants. But using his skill as a physician and years of experience, he had figured out that the drug supply was “sick,” as he described it. Whenever Lever switched his patients’ drugs, he tried to keep the dose the same. This helped to eliminate another X in the equation.
Sometimes, even the same doses of different versions became dangerous, as Lever found when treating Kevin Parnell. Parnell was diagnosed with HCM at age thirty-one. Told it was essentially a “death sentence,” he sought a second opinion at Cleveland Clinic. In 1998, at age thirty-nine, the very sick Parnell found himself under Lever’s care. In 2003, open-heart surgery earned Parnell six years of good health. But as time went on, he took more medication and higher doses of the diuretic furosemide.
By 2012, his health was failing. As Parnell sat with his wife in Lever’s office and described the problems with his swollen legs, Lever immediately suspected a poor-quality generic diuretic. Though Parnell had not bothered to bring the medication, his wife called their daughter and instructed her, “Go check the label on Dad’s bottle.” The pills were made by Ranbaxy. Lever switched the prescription to a version made by the American company Roxane and kept the dose the same. Parnell started taking his new medication right away and immediately started losing fluid. But three days after the switch, he got up in the middle of the night, then passed out and hit his head on the bedside table. He awoke to the sight of three paramedics hovering over him. He was rushed to a local emergency room and was later admitted to Cleveland Clinic. He had suffered ventricular tachycardia—an overwhelming, rapid heartbeat.
Lever suspected that once Parnell changed to a more effective drug, the dose was suddenly too high. The diuretic had rapidly depleted Parnell’s potassium levels, causing his heart to beat irregularly. Parnell finally received the heart transplant he desperately needed. After his surgery, Lever “was in the ICU constantly,” Parnell recalled. “He checked on me at all different hours.” He credited Lever with saving his life. Lever was particularly infuriated by Parnell’s case. He felt certain that drugs of varying effectiveness had harmed Parnell and had reduced the science of medicine to guesswork.
As Lever regularly paged Cleveland Clinic’s pharmacists and alerted colleagues to problematic drugs, he began to reshape the list of drugs the health system used. Cleveland Clinic’s vigilant pharmacists essentially ran a mini-FDA: they sought out bioequivalence data from generic drug makers, investigated the origin of active ingredients, submitted Freedom of Information Act (FOI) requests to the FDA for additional data, and even visited manufacturing plants, all to figure out which drugs the hospital should use and which they shouldn’t. As they heeded Lever’s warnings and reached for a larger set of data points—scouring FDA inspection reports and warning letters and fielding anecdotal reports from doctors—they developed a confidential black list of drugs the hospital would no longer buy, dominated by generic drugs manufactured in India.
Dr. Randall Starling, a member of Cleveland Clinic’s Heart Failure and Cardiac Transplant Medicine Section, was stunned to learn in late 2013 that a generic version of tacrolimus, made by the Indian company Dr. Reddy’s Laboratories, was on the black list. Tacrolimus is a crucial drug for transplant patients, because it suppresses the immune system to prevent organ rejection. The Dr. Reddy’s version was by far the cheapest, but the number of recalls of Dr. Reddy’s drugs had made the health system’s pharmacists uncomfortable.
Over the next six months, Starling worked with his staff to make sure that none of their patients was taking the Dr. Reddy’s version of tacrolimus and that the hospital’s inpatient and outpatient pharmacies no longer carried it. Though he told patients to use only Prograf, the brand-name version of tacrolimus, Starling knew that once his patients left the clinic’s grounds, he could not control what an outside pharmacy might dispense. It did not take long for his fears to be realized.
In October 2014, about eighteen months after his heart transplant, Cedric Brown, a forty-eight-year-old patient, was admitted to the cardiac medicine service with symptoms of acute organ rejection. Brown swore that he had taken every pill. Up until that point, Brown’s post-transplant recovery had been remarkably successful. He had been up and walking after a few days, and left the hospital within two weeks. The Prograf, which suppressed his immune system to prevent rejection, cost about $3,000 a month, and Brown would have to take it for the rest of his life. By the time he was readmitted to the hospital, he’d gained fifty pounds and felt terrible. He was released after a month, but wound up back in the hospital’s intensive care unit within a week. He didn’t know if he was going to die: “I just prayed to God.”
On a Monday morning, Dr. Starling came in to consult on his case. Standing by Brown’s bedside, he asked him, “Did you get a new prescription?”
Brown had brought in his medicine. “Yes, I got something new at Marc’s pharmacy,” he said. It was a different size and color than the usual Prograf drug.
“Gee, I’d like to see that prescription,” said Starling.
Brown said, “Sure. Open that closet, there’s a bag in there with it.”
Starling fished out the bag, and there was the pill bottle of tacrolimus, made by Dr. Reddy’s. “You are never going to take this one again,” said Starling. And Brown never did. Starling set about to educate the other doctors treating him. In time, Brown recovered enough to go back to work part-time, as a driver. His insurance, Medicaid, covered 80 percent of his brand-name Prograf, and a fund at Cleveland Clinic covered the rest.
Doctors at Cleveland Clinic were not alone in their concerns. In October 2013, a pharmacist at the Loma Linda University Medical Center in California reported to the FDA, through its online complaint database Medwatch, that “multiple patients” who used the Dr. Reddy’s tacrolimus had “unpredictable levels leading to inadequate immunosuppression and subsequent transplant failure.” The report from Loma Linda noted, “This has only been seen with the Dr. Reddy’s brand of Prograf.” Tacrolimus was a so-called narrow therapeutic index drug that required precise dosing; minor variations could lead to life-threatening complications. For years, doctors who prescribed other such drugs, for conditions like epilepsy, hypertension, mood, or endocrine disorders, had debated whether certain generics were really interchangeable with the brand, and whether the FDA’s bioequivalence standards allowed too much latitude. Medical societies, including the American Academy of Neurology, the Endocrine Society, and the American Heart Association, opposed a switch to generics without a doctor’s approval.
Facing medical unease over the interchangeability of certain generics, starting in 2010, the FDA’s Office of Generic Drugs began to commission a series of studies of generic narrow-therapeutic index drugs. In 2013, researchers at the University of Cincinnati began a bioequivalence study of generic tacrolimus made by Sandoz and Dr. Reddy’s. They tested the lowest dose of the drugs in healthy volunteers and followed two groups of patients who’d gotten kidney and liver transplants. Their results, which they published in 2017, found the drugs to be bioequivalent to the brand and interchangeable with one another.
But at Cleveland Clinic, Starling and his team felt little reassurance. Months after Cedric Brown’s admission, it happened again—another heart-transplant patient suffering from organ rejection who had taken the Dr. Reddy’s formulation was admitted to the hospital. Several more followed. In studying those cases, the only explanation that Starling’s team could find was an inadequate generic. The Dr. Reddy’s drug became a variable with an unknown impact, in a treatment plan with no room for error. Starling was a doctor who valued control and was accustomed to solutions, but he was left without one. He had no Cleveland Clinic data to point to. But like his colleague Lever, his “index of suspicion” increased.
“I developed a stance of demanding that my patients take brand-name drugs,” Starling explained, “because I didn’t want any variable out there.” Though he knew this created a financial and logistical burden for Cleveland Clinic, its patients, and their insurers, the stakes could not have been higher. An organ transplant is a “huge investment,” said Starling—organs were scarce, and the average cost of transplanting a heart was well over $1 million. “If we’re giving a patient inert medication, it’s a huge failure of the system. An organ has been wasted potentially.”
Both Starling’s and Lever’s patients had served as sentinels. In their fluctuating heart rhythms, Lever sensed that something was awry. It was a problem he couldn’t see and had no way to fix, but he suspected it lay in the distant manufacturing plants where his patients’ drugs were made. In the months that followed, as he took his quest to the FDA and to the news media, his instincts were proven true. “In all my innocence,” he said, “I stumbled on a mess.”