Archive for January 2012
One of the more interesting aspects of having a blog is getting comments from readers. It is always nice to get responses that enjoy the posts or that value my views. Of course, there are also some that suggest I don’t know what I’m talking about. I received one such comment last week which I believe merits a considered response. A couple of weeks ago, I wrote a piece entitled: “Disease Creep or Good Medical Practice?” Essentially, I was challenging the view that the use of statins to lower LDL cholesterol is an example of “disease creep” – the conversion of a risk-factor to an actual disease. Briefly, my view is that high cholesterol, while not a disease itself, is a strong predictor of heart disease in people with multiple cardiovascular risk factors and thus needs to be treated if diet and exercise aren’t sufficient to lower LDL cholesterol to recommended levels.
A few days later, I received a comment that began: “Did none of you receive the message that statin use was associated with the onset of type 2 diabetes?” The writer was referring to recent data from the Women’s Health Initiative (WHI) reported in the online edition of “Archives of Internal Medicine.” The investigation included 153,840 post-menopausal women aged 50 – 79 who were not diabetic at the start of the study. Over the course of the next 6 – 7 years, 10,242 women developed type 2 diabetes. After taking into account factors like lack of exercise and weight gain, it was determined by the WHI researchers that those on statins were 48% more likely to become diabetic than those not on statins at all.
Does this mean that statins can make older women more susceptible to diabetes? That is hard to say. It is possible that statins have effects on muscles and the liver that cause improper glucose metabolism, which could lead to diabetes. However, the WHI study is an observational study, not a prospective one and there are uncontrolled factors in an observational study. For example, women on statins may feel that they are protected from heart disease and may be eating more starches and sugars. Post-menopausal women don’t metabolize sugars well and this may be contributing to the increase in incidences of diabetes seen in this study.
Observational studies can yield wildly different results. Some, like the WHI study suggest negative effects of statins, others show unexpected benefits. It is interesting to note that a recent observational study suggests that taking statins may prevent death from flu. In a study entitled “Association Between Use of Statins and Mortality Among Patients Hospitalized With Laboratory-Confirmed Influenza Virus Infection: A Multistate Study” published in The Journal of Infectious Diseases (January 1, 2012), the authors reviewed the records of 3,043 older adults hospitalized with the flu in 2007/2008. Those who were on statins were 41% less likely to die from the flu over 30 days compared to those who weren’t, independent of age or flu vaccination status. An accompanying editorial to this article stated the dilemma that all observational studies pose: while there is a scientific rationale as to why statins could reduce mortality, the results might also be explained by a “healthy user bias” in that “statin users are more apt to be discriminating users of healthcare.” The editorial emphasized the need for a double-blind, placebo-controlled randomized trial of acute statin therapy in hospitalized statin-naïve, influenza infected patients to provide a definitive answer as to whether statins have protective benefits in flu patients.
Randomized prospective studies do exist for statins in diabetics. One of note is CARDS (Collaborative Atorvastatin Diabetes Study). CARDS was designed as a 6-year study and was made up of 2838 diabetes patients aged 40 – 75 with no history of heart disease. The patients were randomized to receive 10mg of atorvastatin or placebo. Over time, the LDL cholesterol levels of those on placebo remained unchanged (at about 120mg/dL, whereas those on atorvastatin had dropped to 77mg/dL on average. Surprisingly, the Data Safety Monitoring Board (DSMB) halted CARDS after only 4 years when it was found that those on atorvastatin had 35% fewer cardiovascular events, 48% fewer strokes and a death rate lowering of 27%. When the results of CARDS were reported in The Lancet (Vol. 364, 685 – 696, 2004), the authors concluded: “The debate about whether all patients with type 2 diabetes warrant statin treatment should now focus on whether any patients can reliably be identified as being at sufficiently low risk for this safe and efficacious treatment to be withheld.”
Statins are important drugs. But like any drug, they can have risks and need to be used appropriately.
In 2011, the FDA approved 30 New Medical Entities (NMEs) filed either as New Drug Applications (NDAs – small molecules) or Original Biologic License Applications (BLAs – therapeutic biologics). This is the largest number of approvals since 36 new drugs were approved in 2004. Furthermore, on average only 23 new drugs were approved in the previous decade, so the 2011 total is pretty impressive. People are debating whether this represents a turnaround for the biopharmaceutical industry or a one year aberration. Regardless, as good as 2011 was, there is no doubt that the number of new drugs produced by the industry is significantly lower than what it produced in the 1990s.
When people talk about this issue, they always use 1996 as their starting point. In that year, the FDA approved 53 NMEs, an all-time high. However, using the 1996 data as a starting point for a productivity discussion is totally inappropriate as THAT was a one year aberration. In the early 1990s, the U.S. was undergoing a “drug lag,” that is, drugs were being approved more rapidly abroad than in the U.S. As a result, a number of drugs were languishing at the FDA for years before approval. Needless to say, Congress got involved and they realized that the FDA was under-resourced to approve drugs in a timely fashion. To solve this problem, Congress enacted the Prescription Drug User Fee Act (PDUFA), a mechanism whereby charges were levied on pharmaceutical companies for each new drug application filed. The revenues from these “user fees” were used to hire 600 new drug reviewers and support staff. This personnel increase enabled the FDA to work through the backlog of NDAs. The record number of NDA approvals in 1996 is a result of this.
(As an aside, the user fee in 1995 for a full NDA was $208,000. In 2012, the fee is $1,841,500. Considering that fewer NDAs are being filed, given the nine-fold increase in PDUFA user fees, one might wonder why all drugs can’t be approved with a six months review time.)
Nevertheless, the FDA approved 315 new drugs from 1991 – 2000. Thus, despite the progress in technologies for drug discovery and development, as well as the wealth of information that has emanated from the Human Genome Project, the number of new drugs emerging annually has dropped significantly. What may account for this? For one thing, due to industry consolidation, there are fewer companies producing new drug candidates. In 1988, the Pharmaceutical Research and Manufacturing Association (PhRMA) had 42 members. Only 11 of those companies exist today. While there are biotech companies like Amgen that have arisen over this timeframe, there haven’t been nearly enough new companies formed that could make up for this decrease in NDA-producing organizations. Back in 1990, if a new idea for treating cancer arose, 25 different companies would have jumped on it. Given the challenges inherent in R&D, one might assume that 4 or 5 of these companies would have been successful in getting such a drug approved. Things are quite different now. With fewer companies competing, the chances for success drop precipitously. Not only are there then fewer NDAs produced, you also wind up getting fewer entrants in a new class of drugs, which provides fewer choices to patients, physicians and payers.
Another reason for the decrease in NDA output is the result of the higher safety and differentiation hurdles experimental medicines face. For a new drug to be a commercial success these days, it needs to be differentiated from existing therapies. In addition, the FDA is requiring that new compounds are effective in treating the disease itself and not just impact markers of the disease. For example, a new compound that lowers LDL (“bad”) cholesterol might be of value, but the FDA now requires data actually showing that such an agent actually reduces heart attacks in a patient population with cardiovascular disease. In the 1990s, simply lowering LDL was enough to get FDA approval. Outcomes studies, if they were done at all, would be done after the drug was approved and marketed. Now, extensive clinical trials are needed and the new agent not only has to reduce heart attacks and strokes, it also needs to do so better than the generic statins.
The impact on industry productivity as a result of higher differentiation and efficacy hurdles can be seen in the decrease in compounds clearing late stage (Phase 3) clinical studies. In the 1990s, >90% of compounds entering Phase 3 received FDA approval. Data from Arrowsmith (Nature Reviews Drug Discovery Vol. 10, 87, February 2011) suggests that only 50% of compounds entering Phase 3 get FDA approval. Phase 3, which prior to 2000 served to confirm results from early smaller scale clinical trials, now is a major hurdle in determining a drug’s ultimate medical and commercial value. The drugs that emerge from such vigorous trials are likely to be major advances. However, fewer of these occurrences happen now than a decade ago.
For NDA approvals, the 1990s can be viewed as an era when there were many large organizations producing multiple compounds that didn’t need to be differentiated from existing therapy nor show multiple year safety and disease reduction in patients. The rules, however, have changed. The hurdles and costs for new drug development are higher than ever before. As a result, to expect the industry to produce 50 new drugs per year is unrealistic, regardless of technology advances. Thirty is the new 50.
A brief article in The Economic Times (January 2, 2012) entitled “Zydus, Eli Lilly drug discovery deal off” got very little notice, but may represent a new trend. Three years ago, the Indian company Zydus Cadila signed a research pact with Lilly to collaborate on novel drugs to treat cardiometabolic disorders. The deal had the potential to be worth up to $300 million if Zydus Cadila achieved specific milestones and compounds passed key steps in clinical development. While not a lot of information was given for the dissolution of this collaboration, I found the following comment interesting:
“Developing a new drug from scratch is getting more expensive due to increased regulatory scrutiny and high costs of clinical trials. Lowering costs through a partnership with an Indian drug firm was one way of speeding up the process, but the success rate has not been very high.”
Much has been made of the growing investment in R&D being made by Western pharmaceutical companies in places like China and India. Initially, this work was meant as a way to get laboratory work done more cheaply. However, about a decade ago, Western companies turned to their outsourcing partners to tap into their scientific talent and to broaden their search for new medicines. This blended well with the ambitions of these countries to become bigger players in the lucrative global pharmaceutical market. A few years ago, this trend was discussed by Pete Engardio and Ben Rissing in a Business Week story entitled “Big Pharma’s R&D Booster Shot” (June 11, 2008) which focused on India and highlighted the proliferation of deals, such as India’s Glenmark Pharmaceuticals liaison with Lilly and Ranbaxy’s drug-discovery alliance with Merck. The authors acknowledged that, while Indian firms were unlikely to compete in the development of new medicines overnight due to the long drug discovery-development timeline, they were clearly on the path to be eventual competitors.
But the Indian firms are now discovering how difficult this business is. The Zydus Cadila – Lilly example is not unique. Similar disappointing stories can be told for the early diabetes drugs that Dr. Reddy’s Laboratories licensed to Novo Nordisk and Novartis, or Glenmark’s anti-asthmatic compound that was licensed to Forest Laboratories. However, these setbacks shouldn’t be thought of as a failing of India’s nascent pharmaceutical efforts. They just provide further evidence of how challenging drug discovery is, no matter where it is done.
However, there could be at least one concern with India’s emerging drug pipeline. In his article “India Perseveres as Drug Discoverer” (Chemical & Engineering News, October 31st, 2011), Amruthanand Nair shows a table listing the drug candidates in advanced development from several major Indian firms, including Dr. Reddy’s, Glenmark, Piramal Life Sciences and Sun Pharma. Of the 10 compounds listed, many are mechanistically related to agents that are already on the market or they are in classes where previous compounds have had little or no success. It is hard to imagine that any of these agents will be blockbusters without long-term and expensive clinical studies that would differentiate them from current therapies. Do these small companies have the internal resources to carry out such studies? If not, they will need a major pharma partner to do this. And this comes back to the “developing a new drug from scratch” quote above. It is entirely possible that, as happened with the Zydus Cadila – Lilly deal, Western partners will be dissolving these deals unless there is compelling evidence that the compounds generated will meet not just regulatory scrutiny but also the very high hurdle that now exists for compound differentiation.
When it comes to drug discovery, the Indian pharmaceutical industry is still in its infancy. It will be awhile before the “eventual competitor” status will be realized.
The American Heart Association has been monitoring deaths due to cardiovascular disease (CVD) in the U.S. for over a century. While the CVD death rate grew steadily for most of the 20th century, it leveled off and then began to drop somewhat over the past 25 years. Nevertheless, CVD is still the leading cause of death in the U.S. with 600,000 people dying annually, which accounts for more than 25% of all deaths in the country. The direct costs associated with treating heart disease amount to over $80 billion/year and indirect costs attributed to loss of productivity exceed $60 billion/year.
Despite the progress made in moderating the CVD death rate, it is still a major disease. Furthermore, as the obesity epidemic continues in the U.S., recent headway is liable to be counteracted by the increase in obesity, which is already resulting into a concomitant increase in type 2 diabetes, a precursor to heart disease. Even with improvements in diagnosis and treatment, better understanding of risk factors, reductions in smoking, etc., CVD is going to remain a major health problem for decades.
Diet and exercise are keys to staying healthy, not just to ward off heart disease and diabetes but other diseases as well. Yet, there are times when medical treatment becomes a necessary add-on to preventing heart attacks and strokes. It is not a coincidence that the lowering of CVD deaths occured at the advent of statins such as Zocor (simvastatin) and Lipitor (atorvastatin), drugs that lower LDL cholesterol which is a key factor in the formation of atherosclerosis. These medicines have been shown to be both effective and safe for long-term use.
Thus, it is stunning to see an essay like Jeanne Lenzer’s “Disease Creep: How we’re fooled into using more medicine than we need” (December 22, 2011). Lenzer’s views can be summarized in her quote below:
“Elevated cholesterol is not a disease. It doesn’t cause symptoms. It is a risk factor. People with high cholesterol levels are somewhat more likely to develop a heart attack or stroke, but they are at far less risk than individuals who already have cardiovascular disease. This is the definition of disease creep: when pre-conditions or risk-factors are treated as if they are the same as the actual disease state.”
In Lenzer’s utopia, you wouldn’t get a statin until AFTER you have already had a heart attack. The problem is that many first heart attacks are fatal – you don’t get a second chance to go on statin therapy then. She is correct in saying that just having high cholesterol alone does not justify taking a statin to prevent a heart attack or stroke. But CVD risk factors also include male sex, older age, family history of heart disease, post-menopause, smoking, obesity, high blood pressure, diabetes and stress. If a patient presents to a physician with multiple risk factors and if diet and exercise have not been effective in lowering cholesterol levels to those recommended by the American Heart Association, that physician would be remiss if the patient wasn’t prescribed a statin. Waiting for a patient to first have a heart attack or stroke before providing such treatment would be irresponsible.
Lenzen implies in her article that the prophylactic use of statins may only prevent 1 in 50 heart attacks. I don’t necessarily agree with that number, but let’s say that is correct. There are 785,000 first heart attacks/year in the U.S. Even employing Lenzen’s assumptions, the use of statins in the overall treatment paradigm of patients with multiple CVD risk factors would prevent thousands of heart attacks or strokes annually. Now that the most-studied statins like simvastatin and atorvastatin are generic, it would seem like the cost-benefit of statin use to prevent first heart attacks is non-controversial. This isn’t “Disease Creep” – it is simply good medical practice.