Friday, December 13, 2013
"I am throwing up blood," she said.
"For how long?" I asked. "What color is it?"
"Since this morning," Jane said. "It is bright red."
Jane was eighty-four, and during the three years in which I'd been her neurologist she'd been having what she called the "dwindles". First one stroke, then the next, then a series of small strokes, each imprisoning her vibrant mind in her small body. Incapacitated, weighing no more than a hundred pounds but still in control of her thoughts, she was living in her apartment on Fifth Avenue, ten blocks from my office at Park Avenue and East 76th Street.
"Jane," I said, "do you want to go to the hospital?"
"No," she said.
"Alright," I said, "you know this is serious, it means you will bleed to death. If you stay home, you will die."
"I know," Jane said, "I'm done."
And I knew it too. Even if we successfully treated this new trauma, I knew that Jane herself would be further diseased, further decimated. I was relieved. The last three years had been a long and painful struggle for us both.
When I first met Jane in the autumn of 1998, I was struck by her abundant energy and formidable intelligence, a handsome woman fiercely committed to a variety of philanthropic causes, who dressed in tailored twill and drove, at high speed, a hunter green convertible with cream leather seats. I'm still not sure why she chose me to be her neurologist, a young woman not long out of her fellowship training, inexperienced as a practicing physician, low on the totem pole of the medical faculty at Columbia University.
Why me and not one at the department chairs, the hospital bigwigs? Maybe she regarded me as one of her causes, along the lines of an unconventional modern building that she thought deserving of landmark status. Soon after we were introduced, she had said, "You are going to be something special."
Three months later she said, "Can I have a copy of your book?"
"What book?" I said. And then, of course, I wrote one.
Jane had a way of making things happen, saving beautiful old buildings from the wrecking ball by browbeating donors into donating, legislators into legislating. On that Tuesday afternoon in August I hurried over as soon as I could. Sam, Jane's husband, opened the door. He was a slight man, courteous and old-worldly, dressed like a country squire, his gray hair neatly combed back, the lapels of his jacket lying flat against his chest.
"Steve is here," Sam whispered in the foyer, nodding toward the bedroom. "He wants Jane at the hospital immediately. He wants us to call 911."
"And you, Sam?"
"I am with Jane. She's been just as brave as can be." Sam eyes became watery. He cried each time she had a stroke. He cried when he laughed. In the last few months, just the mention of Jane was enough to get Sam crying. But Sam also cried when the stock market hit a new low and when he saw a beautiful piece of art. Tears came easily to Sam, and gloriously, unabashedly, rolled down his rosy cheeks as they must have doing since he was a child.
Steve was Jane's plastic surgeon. This being the Upper East Side of New York City, plastic surgeons were as important as neurologists, often more so. Jane, in her pragmatic, no-nonsense way, had had some "work" done in her later years. She saw no need to defend the mingling of the natural and the artificial. Her ability to combine disparate elements into a serene whole was one of her signature and inexplicably endearing features.
I went into Jane's bedroom, a large, high-ceilinged, draperied Fifth Avenue room overlooking Central Park, airy and light despite the period French furniture. Swirling around Jane in this room now was Marta, the maid, Lulu, Jane's silky, energetic terrier, and, of course, Steve. He was wearing a blue polo shirt and khaki slacks, his dark hair slicked back over a handsome, Slavic face upon which quite some "work" had at some time been done. Steve had the type of preternaturally taut, middle-aged body that only good surgical suctioning achieves.
Seeing the buzzing around Jane lying prostate in bed, I felt myself on the set of the Ars Moriendi, the small medieval manual featuring eleven woodcuts on how to die the Good Death. The scenes are more crowded than the main floor of Barney's at Christmas season, teeming with saints, devils and people. The devils slither around and under the dying person, the Moriens, while the saints hover above and around the bed. Relatives and friends run off with the horses in the stable and the fine wine from the cellar of the soon-to-be-dearly-departed.
In twenty-first century America, there is no such "how to" manual on dying. Nor does our state-of-the-art modern medicine offer much help.
Fact: Seven out of ten Americans wish to die at home, die the Good, the Valid, Death.
Jane abhorred whispering, so Steve and I included Jane in our discussion of the mechanics of her death.
Fact: Seven out of ten Americans die in institutions, intubated, infiltrated, invalidated.
"This is a treatable problem," Steve said.
"Yes," I said, "but she is going to be worse off afterward."
We went back and forth like this for half an hour. Jane and Sam listened intently, following the argument like spectators at a tennis match. In the end, Steve acquiesced even though his suggestions were in line with what my hospital colleagues would recommend. What I was proposing was not the medical norm.
I felt wobbly and unsure. Steve had decades more experience as a physician. What if he was right and I was wrong? What if it was wrong to agree with Jane and let her go, wrong to let her die? I had never helped anyone die before. How would I go about doing this? There were many colleagues I could ask to help me help Jane live, but I couldn't think of anyone to call to help me help Jane die. I could, of course, hand her off to a hospice team, an end-of-life, medical clean-up crew. But this felt like a shirking of responsibility, an abandonment of my patient in her time of need, so that I would be spared the discomfort. The opting out seemed to me cowardly.
I told Sam to alert everyone in the family. The time, I said, had come.
"How long?" Jane asked me. I didn't know how to answer the question.
I looked at Lulu, at her perky, twitching ears and her eyes shimmering in moist health. She lay on the pillow next to Jane, occasionally swishing her little tail the way dogs do when they are being looked at. I was trying to see if she knew, as dogs often know things that people don't, but Lulu just kept on peering mutely back at me.
"It could be in a few days," I said. "We will make sure that you stay comfortable."
We? Who was the "we"? In my tentative foray into this new territory of helping someone die, I was waving the flag of the larger medical establishment to reduce my vulnerability and make me feel and seem more authoritative.
"Okay," Jane said, squeezing my hand. The firm way Jane grasped my hand said in so many words that she was putting her trust in the specifics of me, not in the nebulous "we" of established medicine.
Steve's disapproval was like a thick fog in the room. I kept my posture erect as I left, walking very slowly, measuring the pace of my feet and breathing in and out. Once off stage, I sprinted to the elevator as if I was being chased.
Thursday, December 12, 2013
It was almost 15 years ago that I set my personal record, since equaled but never exceeded, of 18 patients seen in the course of a single afternoon. What I remember are not the details of that particular day — all days are hectic for one reason or another — but a conversation I had at a party a short time later.
With nothing much else to discuss with another guest, also a doctor, I modestly mentioned my record. He smiled indulgently and replied that he averaged more like 60 primary-care visits a day.
For some time afterward, I compulsively redid the math (he spent around eight minutes per patient, hour after hour) and idly fantasized about his office, which clearly included a small army of medical assistants, other capable souls to handle phone, fax and paperwork, and extraordinary data-retrieval abilities for labs and tests. And a lot of agreeably healthy patients. Or maybe they were too sick to talk much. Did the man eat? I wondered. Did he type? Perhaps a court reporter stood by.
Of course, when I heard a few years later that the state had closed his office and stripped him of his medical license, a whole new set of images came to mind.
I am hardly alone in these fantasies: Every medical administrator in the country has the same obsession with physician productivity, from the happy daydreams of smooth high-throttle operation to the nightmares of wipeout if the machine spins out of control.
For some medical metrics, the sky's the limit — we aim for zero complication rates, 100 percent adherence to various required annual screens. But performance statistics always come with a natural cap. The cheetah maxes out at 60 miles per hour or so, and no American cow has yet provided more than 7,000 gallons of milk a year. How fast, then, can a domestic physician move? How much health can one produce?
Everybody approaches the math a little differently, wielding different units of measurement. Some think only in dollars, with yearly income goals that are self-imposed or set by an employer. The less said about them the better.
A more legitimate unit of productivity is the number of patients in a doctor's practice — the patient panel, to use the technical term. American primary care providers have long been estimated to work smoothly if they care for about 2,300 patients each, at an average rate of 3.3 visits a year, which works out to 25 visits a day, including Saturdays.
Some authorities now worry this number may be much too high. Even without factoring in the exploding paperwork, the list of tasks each patient brings to the office is growing with body mass. The average American general practitioner is now estimated to worry about a total of 417 cases of uncontrolled cholesterol, 312 cases of uncontrolled high blood pressure, 91 cases of uncontrolled diabetes, a few hundred of these same problems under good control and then a variety of other chronic illness in various states of disarray.
Wednesday, December 11, 2013
Curb ‘weekend effect’ at hospitals to make Saturday or Sunday admittance less risky for patients: study
For 20 years, the evidence has piled up in study after study: Patients admitted to hospital on the weekend are more likely to die or have poor outcomes, as reduced staff levels cause delays in tests and treatment.
Yet the health-care system has done little or nothing to address the problem, say two Canadian doctors who have just published a call to action for hospitals to curb the so-called "weekend effect."
The problem is complex and the solutions would be no less so, but it is time for health care to "restructure the hospital work week" and make Saturday and Sunday just as safe for patients as any week day, they said.
Other doctors who have studied the problem echoed their appeal, suggesting that even relatively minor, inexpensive changes could save lives.
"It really is a pervasive thing that has happened in multiple aspects of medicine," said Dr. Chaim Bell, co-author of the new article in the journal BMJ Quality and Safety. "The problem has been trying to get a call to action. It's hard to change an ingrained approach to care and organizational structure."
One prominent patient advocate suggested Tuesday that achieving an overhaul of scheduling that resulted in more shifts on Saturday and Sunday might, in fact, be all but impossible.
"Health care is delivered primarily by middle-class people," said Sholom Glouberman, spokesman for the group Patients Canada. "Doctors tend to be upper middle-class people, and they tend to arrange their lives so they don't work on the weekend."
A representative for many of the country's hospitals, meanwhile, said he welcomed Dr. Bell's editorial, while calling for more study.
"This research is important," said Anthony Dale, interim CEO of the Ontario Hospital Association. "A first step will be to fully understand the facts behind the findings, such as prevalence and contributing factors, so that they can be properly examined and addressed."
In the meantime, patients should not fear going to hospital on the weekend, but ask caregivers if any delay in testing or therapy is safe, experts say.
Dr. Bell, an internal-medicine specialist at Toronto's Mount Sinai Hospital, along with colleague Dr. Donald Redelmeier of the city's Sunnybrook Health Sciences Centre, have been among the pioneers of research in the area. Their 2001 review of Ontario hospital data found patients admitted on the weekend were significantly more likely to die of 23 common, potentially fatal conditions, and no less likely to die from other ailments.
The medical literature over the past several years has been replete with similar papers, bearing titles like "Don't get sick on the weekend" and "Weekends: a dangerous time for having a stroke?"
A U.S. study published this year found a relatively modest 3% increase in mortality for patients admitted to hospital on the weekend. But that would translate into "thousands and thousands" of potentially preventable deaths, said Dr. Adam Sharp, a University of Michigan emergency-medicine doctor and co-author of the research.
"The day of the week should not predict whether you die or not, no matter how small the difference," he said.
A 2007 Canadian study suggested that stroke patients who landed in hospital over the weekend were 14% more likely to end up dead.
"If you look at how many cars are in the parking lot on weekends, versus the weekdays, you will realize the basic difference," said Dr. Gustavo Saposnik, a neurologist at Toronto's St. Michael's Hospital and the lead author. "And that is obviously translated into the care, the number of providers available to provide excellent care."
Researchers stress that they can only speculate about the reasons for the weekend effect. As well as fewer staff and the resulting delays, it could also be a function of less-experienced, more junior nurses and doctors being on duty, they say.
And the solutions are not as simple as they might appear. A wholesale scheduling change that allowed hospitals to function at the same level on the weekend as during the week would likely be prohibitively expensive and unnecessary, said Dr. Sharp.
It will be important to focus on specific areas where weekend care is proven to lead to poor patient outcomes, and address them directly, Dr. Bell said.
In the care of stroke patients, that does not necessarily mean having more people on duty, said Dr. Saposnik. Much could be done by ensuring weekend staff are trained to perform key tasks, he said. Those include swallowing assessments designed to ensure patients do not aspirate food into their lungs — a potential source of infection — and simply getting stroke sufferers to walk around, staving off potentially dangerous blood clots in their legs.
Mr. Glouberman suggested, however, that energy should be focused, not on beefing up hospital services on the weekend, but improving community-based care for people with multiple illnesses, partly by training more multi-skilled geriatricians.
That way, such patients would be less likely to have to be admitted to the hospital — on any day of the week, he said.
Tuesday, December 10, 2013
An effective eye drug is available for $50. But many doctors choose a $2,000 alternative. - The Washington Post
But one holds a clear price advantage.
Avastin costs about $50 per injection.
Lucentis costs about $2,000 per injection.
Doctors choose the more expensive drug more than half a million times every year, a choice that costs the Medicare program, the largest single customer, an extra $1 billion or more annually.
Spending that much may make little sense for a country burdened by ever-
rising health bills, but as is often the case in American health care, there is a certain economic logic: Doctors and drugmakers profit when more-costly treatments are adopted.
Genentech, a division of the Roche Group, makes both products but reaps far more profit when it sells the more expensive drug. Although Lucentis is about 40 times as expensive as Avastin to buy, the cost of producing the two drugs is similar, according to scientists familiar with the drugs and the industry.
Doctors, meanwhile, may benefit when they choose the more expensive drug. Under Medicare repayment rules for drugs given by physicians, they are reimbursed for the average price of the drug plus 6 percent. That means a drug with a higher price may be easier to sell to doctors than a cheaper one. In addition, Genentech offers rebates to doctors who use large volumes of the more expensive drug.
"Genentech continues to maintain that Lucentis is the most appropriate medicine," the company said in a statement, adding that it costs "significantly" more to make and is tailored for use in the eye. The drug "has made an immense impact."
Many ophthalmologists, however, are skeptical that it provides any added value over the cheaper alternative.
"Lucentis is Avastin — it's the same damn molecule with a few cosmetic changes," said J. Gregory Rosenthal, a Toledo ophthalmologist who, outraged by the price, co-founded a group called Physicians for Clinical Responsibility to protest its use. "Yet Americans are paying a billion dollars every year for no good reason — unless you count making Genentech rich."
The story of Genentech's two drugs, Lucentis and Avastin, began with a scientific marvel — a breakthrough in biology that, thanks to the vast budgets of U.S. entitlement programs, has produced enormous financial returns.
Those profits have yielded benefits. By paying for such drugs without regard to cost, the Medicare system has helped stimulate investment in medical research that contributes to the development of more lifesaving technologies.
But the flow of cash also pushes up the health-care costs that are projected to deplete federal budgets. For while Genentech has aggressively marketed the more expensive drug and sought to restrict the use of the cheaper one, critics say, Medicare has been powerless to do anything but pay up.
That's because over the past seven years, despite pleas from the Food and Drug Administration and doctors groups, Genentech has maintained the barriers that make it harder for doctors to use the cheaper drug.
Avastin was not originally intended for use in the eye, and the company has refused encouragement from the FDA to seek official approval for using it to treat eye ailments, according to unpublished internal FDA documents. This forces doctors to use it "off-
label," or in ways not specified on the medicine's label.
The company also packages the drug, which was approved for cancer in 2004, in doses far too big for use in ophthalmology, meaning that the drugs must be repackaged by other companies for use in the eye, raising the risk of contamination.
Genentech has argued that Avastin may pose a greater danger of severe side effects than does Lucentis, although independent scientists say such worries are unsupported by the six trials that have been conducted.
In a statement, the company said that it has not sought FDA approval of the cheaper drug for use in the eye because it has already developed one drug for the ailment known as wet age-
related macular degeneration, or wet AMD.
"Genentech continues to maintain that Lucentis is the most appropriate medicine for wet AMD as supported by clinical and other scientific data," the statement said.
"We specifically designed Lucentis for use in the eye and to clear quickly from the bloodstream after leaving the eye to potentially minimize side effects," the statement said. "The two medicines were designed for different purposes and, we believe, may have different systemic and ocular safety profiles when used in the eye."
Genentech defended its pricing by noting that the Roche Group spends $9 billion annually on research and development.
"The price of Lucentis supports the research and development of new potential medicines, including the 92 percent of drugs that never make it to patients," the company said. "We re-invest a larger portion of our revenue into clinical research than most pharmaceutical companies. Genentech believes it is in the best interest of patients to continue to focus our efforts in ophthalmology on discovering and developing new potential medicines for other serious diseases of the eye."
Most doctors, however, prefer to use the cheaper drug. Despite the company's position, U.S. doctors have been using Avastin in about 56 percent of such cases, according to Medicare data analyzed by The Washington Post. In the most recent survey by the American Society of Retinal Specialists, about 61 percent of doctors preferred using Avastin for macular degeneration, with the rest of the market split between Lucentis and Eylea, a new drug made by Regeneron that is almost as expensive as Lucentis.
Because so many doctors continue to use Lucentis, Genentech has rung up more than $1 billion in U.S. sales of the drug for four years running. Roughly 80 percent of U.S. sales are paid for by Medicare and its beneficiaries.
The rising cost of U.S. entitlement programs such as Medicare has prompted outrage in Congress, but it is Congress that has made it difficult in this case and others for Medicare to limit such expenses.
To begin with, the Medicare agency is blocked from seeking better drug prices by negotiating directly with the drug companies, as health agencies in other countries do. Authorities in Britain, for example, have negotiated a price of about $1,100 per dose of Lucentis, and in the Netherlands a dose sells for about $1,300.
Moreover, in cases in which two equivalent options are available, such as Lucentis and Avastin, Medicare is forbidden from restricting payment to the amount of the less costly alternative. After it sought to do so in 2009, a federal appeals court said it lacked that authority.
It's often difficult, of course, to know when two drugs are equivalent. When the debate over the two drugs and their pricing erupted more than six years ago, Genentech asserted that its more expensive new drug was superior. At the time, it was hard to show otherwise. No one had tested them in side-by-side comparisons.
Since then, the six randomized clinical trials involving more than 3,000 patients have found the drugs to be largely equivalent.
Yet in 2012, the Medicare program and its beneficiaries spent $1.2 billion on Lucentis, according to The Post's analysis of Medicare data.
Medicare officials said they have no choice but to pay the bill when a doctor prefers to use Lucentis.
"We do not have the authority to dictate treatment based on cost," Tami Holzman, a spokeswoman for the Centers for Medicare and Medicaid Services, said in a statement. "Under current law, Medicare must cover treatment that is deemed reasonable and medically necessary by a physician or other provider."
Pharmaceutical firms argue that this is the way it should be.
The industry's main lobbying group, known as PhRMA, opposes allowing the government to negotiate prices with companies — a process it calls "price controls" — and similarly opposes attempts by Medicare to pursue a policy of paying only for the least costly alternative.
The industry has spent more than any other in the United States to have its voice heard in Washington. Over the past 15 years, the pharmaceutical industry has far outstripped any other in the money it has devoted to lobbying, according to data from the Center for Responsive Politics. Drug companies spent a total of $2.7 billion over that time.
"Proposals to change this system by imposing price controls or only giving patients access to treatments deemed the 'least costly alternative' by Medicare would have severe unintended consequences," Matthew Bennett, a senior vice president at PhRMA, said in a statement.
Such proposals could discourage medical progress, he said. Moreover, because every patient responds differently to a treatment, it may be difficult for the government to set rules for coverage.
"The cheapest option on average is not always the best option for many patients," he said.
What's the right price for a miracle?
Every year, about 200,000 people in North America are diagnosed with wet age-related macular degeneration, a chronic disease characterized by abnormal blood vessels that leak blood or fluid into the retina.
Sufferers lose clarity in the center of their field of vision, and among older people it has long been the leading cause of blindness.
Then came Avastin and Lucentis.
Both are the outgrowth of pioneering work done by Napoleone Ferrara, a Sicily-born molecular biologist.
Ferrara studied at the University of California at San Francisco and joined Genentech in 1988. First assigned to the company's efforts to develop a hormone called Relaxin, Ferrara devoted his discretionary research to a theory that blood vessel growth could cause cancer and other illnesses.
Over several years, Ferrara and his collaborators identified a protein called VEGF that causes blood vessel growth. They then linked that protein to cancer and macular degeneration. Finally, they developed an "anti-VEGF" drug that would attack VEGF, halting the harmful blood vessel growth.
The first anti-VEGF drug was Avastin, which won approval from the FDA in 2004 for the treatment of colorectal cancer.
Lucentis followed. It is a stripped-down version of the same molecule, and it can likewise attack VEGF and bind more closely to it. It won FDA approval in 2006.
"People weren't sure that VEGF would prove particularly important, but sometimes in science, you just follow your own ideas," said Ferrara, now a distinguished professor at the University of California at San Diego School of Medicine. "The magnitude of the benefit of these drugs far exceeded our expectations."
The company spent almost $1.4 billion on the development of Lucentis, which included 18 clinical trials, a Genentech vice president testified to Congress in 2011.
The company appears to have recovered those costs and quite a bit more.
In the first 2 1/2 years, it sold $2.1 billion worth of Lucentis in the United States alone. Another Swiss company, Novartis, in partnership with Genentech, sells billions more overseas.
Much of that is profit.
The company will not disclose how much it costs to manufacture a dose of Lucentis, saying only that it costs "significantly" more to make than Avastin. But scientists knowledgeable about manufacturing drugs of this kind say that the costs of making Lucentis are not much different from those of making Avastin.
Indeed, some scientists said that some aspects of Lucentis make it cheaper to produce.
The Avastin process begins with growing a culture from mammalian cells taken from the ovary of the Chinese hamster.
The Lucentis process begins with growing cultures of the common bacteria E. coli, and these are easier to produce.
The subsequent purification process with bacteria may be more complicated, but "production in bacteria is cheaper than in mammalian cells for several reasons," said Hervé Watier, a medical professor at the University of Tours in France who has studied the drugs.
While there are some "drawbacks" to the bacteria production method, Watier said, "the financial result still remains in favor of bacteria."
"I think the difference in cost in producing them is very modest. They cost almost the same, from what I can tell," Ferrara said.
If so, Genentech is making a lot on each dose. The manufacturing costs may account for 10 percent or less of the price of a Lucentis dose, according to a conservative calculation generated with industry experts.
The company declined to reveal how much it is making from Lucentis above the drug's manufacturing costs.
"Lucentis and Avastin are not the same medicine and should not be treated, nor represented, as if they are," the company said in a statement.
After the development of Lucentis in the early 2000s, it was the only drug known to have such effects.
It seemed to be in a class by itself and seemed poised to win even more in sales than it gathers today.
But then Philip J. Rosenfeld, a Miami ophthalmologist, made a discovery.
Rosenfeld was lead investigator on some of the Lucentis trials that Genentech had conducted, and he recognized how effective it could be.
After reading the research that some Genentech scientists had published, he realized that Avastin and Lucentis were derived from the same antibody and thus were functionally equivalent.
"I realized they would perform in the same way," he said.
Under a university-approved research program, he'd also learned that Avastin, injected into a patient's arm as is done with cancer patients, had the same effects as Lucentis. The trouble was, since the Avastin was going into the entire body, a large dose was needed, and that could produce dangerous side effects. He calculated that a much smaller dose injected into the eye would be just as effective as Lucentis.
In May 2005, Rosenfeld had a patient who was quickly losing her vision. A retired nurse in her 60s, she'd lost the use of one eye already, and none of the available remedies could slow the disease's progression.
Rosenfeld knew that Lucentis could help her, but it would be another year or more before the FDA would approve it.
With the patient's permission, he injected her eye with a small dose of Avastin — one milligram — and ordered her back the next week.
"We were astounded by the results," he said.
The billion-dollar drug Lucentis was about to be beaten to market, and by one of Genentech's own products.
In July 2005, Genentech held what amounted to a coming-out party for its new drug.
At the annual meeting of the American Society of Retinal Specialists, the company presented several detailed studies showing how effective it was in treating macular degeneration. With hundreds of ophthalmologists crowded into the room, speakers for Genentech described the marvel of Lucentis.
"Our jaws were on the floor," recalled Daniel F. Martin, chairman of the Cole Eye Institute at the Cleveland Clinic.
Right after, Rosenfeld presented his Avastin experiment on one patient.
"Phil showed one case report — no animal studies, no randomized trials," Martin said. "But after this meeting, every ophthalmologist on the planet was injecting it. The therapeutic effect was so powerful."
Because Lucentis had yet to win FDA approval and couldn't be sold, ophthalmologists quickly embraced Avastin, which had been approved the year before, albeit as a cancer remedy.
When Lucentis did go on sale, Genentech's blockbuster drug already had a competitor. How could the company convince doctors and hospitals that Lucentis had any major advantage over Avastin?
Over and over again, it sought to discourage the use of Avastin by raising concerns about its safety.
They told doctors that Avastin was not approved by the FDA for use in the eye — Lucentis was. They reminded doctors that if the repackaging firms cutting Avastin into smaller doses were careless, infection could be introduced. And despite the lack of conclusive evidence on the point, they said that Avastin patients might suffer more adverse events than Lucentis patients.
Sometimes, senior FDA officials said, these warnings stretched the truth.
In October 2007, the company announced a move that would severely restrict the supply of Avastin for ophthalmology: It would no longer sell the drug to the repackaging firms that were cutting it into eye-appropriate doses.
The company's president of product development at that time, Susan Desmond-Hellmann, explained in a letter that Lucentis was already available. Moreover, she said that during a routine FDA inspection of the company's Avastin manufacturing facility, "concerns were raised by inspectors related to the ongoing ocular use of Avastin because it is not designed, manufactured or approved for this use."
An FDA ophthalmology official, Wiley A. Chambers, told colleagues that the company had misconstrued the agency's position.
That routine FDA inspection at a Genentech plant, Chambers told his colleagues, was unrelated to the intrinsic safety of Avastin in ophthalmology. Instead, it showed that Avastin had been contaminated by glass particles, a danger that could have harmed cancer patients or eye patients.
"Genentech has found a way to blame FDA for their decision to limit the distribution of Avastin," Wiley wrote to colleagues in an e-mail. "The manufacturing problem at their facility that resulted in glass in their product would be an issue for either the on-label oncology indications or the off-label ophthalmology indications."
Genentech said in a statement: "We have never sought to restrict the ability of physicians to prescribe Avastin as they see fit for their patients. . . . Genentech did not blame the FDA and took the decision independently."
Eventually, after ophthalmologists and their professional societies strenuously objected to Genentech's move to limit Avastin sales — they even threatened lawsuits to make sure the flow of Avastin continued — Genentech backed down and continued to provide the drug to the repackaging firms.
About the same time, Genentech asked the FDA for permission to change the Avastin label to instruct doctors that it was not to be used for eyes. The FDA said there was no evidence to support such a change to the label.
The FDA believed "there was no safety-related basis adequately justifying that labeling change," according to an internal agency e-mail, and the label was not changed.
Today, millions of doses of Avastin have been administered successfully. Six randomized clinical trials around the world, beginning with one called Comparison of AMD Treatments Trials, have found its effectiveness equivalent to that of Lucentis. After the CATT study, the National Institutes of Health issued a news release headlined, "Study finds Avastin and Lucentis are equally effective in treating age-related macular degeneration."
The effort was funded by the NIH because Genentech had refused to test the drugs itself and, in a break from industry custom, had refused to provide the drugs to government researchers. An internal company document described the strategy of not performing a test or contributing the drugs as "in the interests of shareholders and the interests of patients," according to a Senate Aging Committee investigation memo from 2008.
Because it had developed Lucentis, the company said, "there was no need to invest substantial resources and years of clinical development to explore the safety and efficacy of another medicine."
Since the CATT study, five more head-to-head trials have been conducted. They also found Avastin just as effective as Lucentis.
"There have now been six randomized clinical studies that show no difference in the major areas of safety concern — deaths, heart attacks and stroke," said Martin, the Cleveland Clinic doctor who also led the CATT trial.
Indeed, Genentech has acknowledged that the drugs are similarly effective. But the company has argued that Avastin may be dangerous when used in eyes.
"The emerging data consistently show differences in safety — particularly in systemic serious adverse events — between Lucentis and Avastin," Anthony P. Adamis, global head of ophthalmology at Genentech, said in an interview.
These differences are "biologically plausible," Adamis said, because studies have shown that Avastin remains in the blood longer.
The main basis for Genentech's safety argument is a finding in the CATT trial that has not reappeared in any of the following five trials and that some scientists involved regard mainly as a curiosity.
The incidence of what are known as serious adverse events — a catchall category that includes hospitalizations for any reason — was slightly higher in the Avastin group: 40 percent vs. 32 percent. The adverse events included broken bones and urinary tract infections.
"The majority of the adverse events would be difficult to imagine being caused by the drug," Martin said. Martin noted that while small, probably random effects favored Lucentis in some cases and in others they favored Avastin. Neither should be viewed as conclusively related to the drug, he said.
It is very difficult for such trials to detect differences in rare safety events. To do so, a trial might need more than 10,000 patients. Running a trial of that size could cost billions of dollars.
To look for effects in large numbers of patients, researchers often turn to Medicare claims records, examining how patients fared on the treatments in question. It is this kind of review that Lesley H. Curtis, a Duke University medical professor, performed, looking at 146,000 patient claims.
After fully adjusting for patient and provider characteristics, Curtis and her colleagues found that there was no difference in the safety profiles in the drugs.
"In conclusion, we found no evidence of increased risks of mortality, myocardial infarction, bleeding, or stroke," their research paper said.
The other danger to using Avastin, however, has attracted a lot of publicity in recent years.
The fact that the drug needs to be repackaged into smaller doses introduces an element of risk because it opens the possibility that the drug could be tainted during the repackaging process. (Genentech says because the FDA has not approved it for use in the eye, the company cannot legally distribute Avastin in doses appropriate for the eye.
Indeed, in three cases that made the news — in South Florida, Nashville and Los Angeles — just such a problem has arisen. Several patients reportedly suffered vision loss as a result.
"I've never used Avastin because of the potential for contamination," Warren L. Herron Jr., a Pensacola, Fla., ophthalmologist, said after a morning in which he did 11 eye injections. "Is it a rare thing? Yes, it's a rare thing. But I can't stand the idea of ever telling my patients that they can no longer see because I used a tainted drug.
"Besides," he said, "I don't think the extra money being spent for Lucentis is totally wasted because it's going into research and development."
But as Herron noted, the likelihood of contamination is negligible. Globally, hundreds of thousands of injections are doled out every year without trouble, making the risk of contamination in repackaging smaller than the risks that doctors routinely ignore when deciding on a treatment.
Whether a patient gets Avastin, Lucentis or the new drug Eylea depends on an array of factors. Some doctors use only one of the drugs; some let their patients choose; many decisions are guided by whether the patient's insurance covers the entire cost or just a portion; and some doctors may consider how much they earn with each drug.
John Thompson, a Baltimore ophthalmologist who is president of the American Society of Retinal Specialists, noted that most doctors use Avastin and that even more would do so if the company sought FDA approval for using it in eyes and packaged it in appropriate doses.
"If Genentech decided to get FDA approval and make Avastin available in small quantities for the eye," he said, "the American Society of Retinal Specialists would applaud."