Friday, September 01, 2006

Case-Study: "The Rise and Fall of Natrecor"

Health Affairs recently published an article from the Division of Pharmacoepidemiology and Pharmacoeconomics at Harvard about the "rise and fall" or Natrecor [nesiritide], manufactured by Scios, as an example of what is currently wrong with how drugs are researched, marketed, and used. [Kesselheim AS, Fischer MA, Avorn J. The rise and fall of Natrecor for congestive heart failure: implications for drug policy. Health Affairs 2006; 24: 1095-1102.] Note that Scios was merged into Johnson & Johnson in 2003. We posted last year about problems with how Natrecor.

Kesselheim et al suggested:

Pre-Marketing Clinical Trials Are Not Designed to Provide Enough Data to Make Informed Clinical Decisions - Particular problems with the Natrecor trials were that they:
  • Emphasized assessment of "intermediate measures," like laboratory test results, which may not predict the clinical effects of drugs. Trials of Natrecor were designed to measure pulmonary capillary wedge pressure (PCWP), a measurement of how well the heart is pumping, but not outcomes more meaningful to patients, like recurrent hospitalization or death.
  • Excluded many of the patients who might be candidates for the drug in usual practice. In particular, trials excluded patients with diseases that commonly co-exist with heart failure, or patients taking drugs commonly used to treat heart failure.
  • Were too small to have statistical power to assess clinically important outcomes, especially adverse effects. The less common an outcome, the more patients must be enrolled in a trial to determine whether differences in outcome rates across groups receiving different medications were due to the medications, or to chance alone. Some trials of Natrecor found more kidney problems or a higher death rate in patients who received that drug, but the authors dismissed these as "not statistically different," i.e., due to chance alone.
  • Failed to report about possible adverse effects of drugs in an organized way. For example, published papers about Natrecor sometimes omitted mention of kidney problems that might have been due to the drug.
Marketing Encouraged Uses of the Drug Not Supported by Clinical Research - Particular instances cited by Kesselheim et al were:
  • Print advertising in medical journals used graphics that suggested that Natrecor was a general treatment for CHF, "the heart's response to CHF is now recombinant therapy," rather than one that had only been tested in narrow circumstance.
  • Journal articles advocated Natrecor for uses not supported by trial evidence. Kesselheim et al cited articles, "many underwritten by the manufacturer," that suggested using Natrecor as first-line therapy for CHF, as a chronic therapy for stable patients with CHF, and as first-line therapy for acute CHF in the Emergency Department. No trials had been done on the drug in these applications.
  • The manufacturer of Natrecor, Scios, provided materials to physicians explaining how they could more easily bill for Natrecor given as an intravenous infusion to outpatients.
  • The manufacturer supported continuing medical education (CME) activities that promoted Natrecor as a first-line therapy.
Without Organized Post-Marketing Surveillance, Adverse Effects May Not Be Quickly Appreciated - Kesselheim et al asserted "after a drug is approved, there is no organized post-marketing surveillance system in place." Perhaps the lack of such a system accounted for the four year time span from the approval of Natrecor to the appreciation of its hazards.
Natrecor sales were running $400 million in the year before its short-comings became clearer. This case-study illustrates how the current imbalance between the pharmaceutical industry and its regulators leads to the use of expensive treatments whose benefits may not out-weigh their harms.
This week, the New England Journal of Medicine published two articles about how primary care is imploding, particularly because of excess demands that far exceed reimbursement. (Commentary from Medical Rants here and here, and the articles are here and here.)
Maybe if pharmaceuticals (and devices, and many other health care goods and services) were priced more in proportion to their net benefit to patients, we could afford to pay doctors to take enough time to actually take care of all their patients' medical problems. Better evidence about the benefits and harms of goods and services to patients, more honest marketing, and more stringent regulation would help.
Kesselheim et al advocated a number of possible solutions at the policy level. There are some things physicians can do meanwhile:
  • Learn enough about evidence-based medicine to understand the need for skepticism when applying the results of clinical studies to real patients.
  • Become more skeptical of academic review and opinion articles financed by those with vested interests in the products or services advocated.
  • Of course, do not forget that the point of marketing is to get physicians to prescribe or use particular products, not to provide a disinterested review of the evidence.

No comments: