to act promptly to suspend approval of an NDA temporarily, and thereby re move the drug from the market, if it represents an "imminent hazard" to the public health. Once having suspended approval, the Secretary must provide the manufacturer with an expedited hearing on whether the drug should be permanently removed from the market. This special authority is vested solely in the Secretary, and may not be delegated.

The summary suspension procedure provides a critical procedural tool to carry out the obligation of this Department and of FDA to protect the public health and safety. Rapid action may be necessary if scientific data raise serious new questions concerning the safety of the drug. If new evidence or further and more careful analysis of existing evidence indicates that a life-threatening or other serious risk is present, the summary suspension procedure allows the Secretary to end promptly this serious risk. The summary procedure does not eliminate the need to conduct a full administrative proceeding to arrive at a final and conclusive judgment as to whether the drug should be permanently removed from the market.

B. Criteria for Suspension

In my 1977 order suspending the NDA's for phenformin under the "imminent hazard" provisions of the Act, I examined at length the text of section 505(e). the legislative history of the suspension provision, and pertinent court decisions. In re New Drug Applications for Phenformin, Order of the Secretary Suspending Approval, pp. 24-35 (DHEW July 15, 1977). I there concluded that the following factors should be weighed in determining whether approval of a new drug application should be suspended on the ground that continued use of the drug will constitute an imminent hazard to the public health :

1. The severity of the harm that could be caused by the drug during the completion of customary administrative proceedings to withdraw the drug from the general market.

2. The likelihood that the drug will cause such harm to users while the administrative process is being completed.

3. The risk to patients currently taking the drug that might be occasioned by the immediate removal of the drug from the market, taking into account the availability of other therapies and the steps necessary for patients to adjust to these other therapies.

4. The likelihood that, after the customary administrative process is completed, the drug will be withdrawn from the general market.

5. The availability of other approaches to protect the public health.

These criteria were reviewed and upheld in Forsham v. Califano, 442 F. Supp. 203 (D.D.C. 1977).

V. Evaluation of propoxyphene under the criteria for suspension

In analyzing the record in this matter, I have been guided by the expert advice and opinions provided by FDA. In assessing and weighing the evidence, I have recognized that the record of a full evidentiary hearing is not before me.

Under the criteria set forth in part IV above, I am not persuaded that suspension of the propoxyphene NDA's should be ordered at this time. Although I am trouble by the evidence that propoxyphene carries life-threatening risks and is of limited efficacy, I believe that the standards for summary removal of a drug from the market have not been met by the evidence now before me. Therefore, I am denying for the present the HRG petition to declare propoxyphene an imminent hazard.

Nevertheless, because of my concerns about propoxyphene-associated deaths, I have ordered that several steps be taken to minimize as rapidly as possible avoidable harm from the drug and to gather further information on its risks and benefits.

I have directed the Commissioner to have FDA complete expeditiously a comprehensive review of all available information concerning propoxyphene to determine whether the various products containing the drug meet the safety and efficacy requirements of the Act and, thus, whether proceedings should be begun to withdraw the new drug applications for any or all of those products. In the course of this review, FDA will hold a public hearing to receive information and views on the continued marketing of propoxyphene. This hearing is scheduled for April 6, 1979. If at any time during this review evidence appears suggesting that

propoxyphene meets the criteria for suspension, FDA will immediately submit it to me. I will then consider, in light of that evidence, whether to suspend any or all of the NDA's for propoxyphene products.

Three other steps, described below, will provide information to physicians, dentists, pharmacists, and the general public, in order to increase awareness of the risks of propoxyphene, and may result in the imposition of additional restrictions on the production and distribution of the drug under the Controlled Substances Act.

A. Severity and Likelihood of Harm to the Public Health

The principal harm from propoxyphene is death. As HRG points out, propoxyphene is associated with a significant number of deaths. In 1977, the DAWN system reported 607 propoxyphene related deaths, more than those associated with any other prescription drug.

The DAWN data provide, however, only a very rough basis for estimating the true number of deaths that may be caused by use of propoxyphene. The DAWN reports include all deaths in which propoxyphene is found in the bloodstream of the deceased. In some of these cases, propoxyphene, particularly in conjunction with alcohol or a tranquilizer, may have caused the death. On the other hand, if propoxyphene happened to be found in the blood of a person who died in an unrelated car accident, that case would be reported in the DAWN statistics as a propoxyphene-associated death. The DAWN statistics also do not reflect all of the deaths in the country, but include only deaths in 24 major cities, covering slightly over 30% of the total U.S. population. Thus, it is likely that additional deaths associated with propoxyphene are occurring in areas which are outside the DAWN reporting system.

The absolute number of deaths must be viewed in perspective with the actual consumption of the drug. Propoxyphene is very widely used; last year, about 31 million out-patient prescriptions were filled, and additional quantities of propoxyphene were used in hospitals, clinics, and physicians' offices. The ratio of propoxyphene-associated deaths (i.e., the number of times the drug is mentioned in coroners' reports included within the DAWN system) to dispensed out-patient prescriptions is lower than that for the barbiturates, the non-barbiturate sedative-hypnotics, amitriptyline, doxepin, and pentazocine. In fact, propoxyphene now ranks 12th out of 27 drugs in ratio of drug-associated deaths to dispensed prescriptions.

The reason for these deaths has long been thought to be suicide. Undoubtedly this motivation accounts for a significant proportion of the deaths. In its petition, HRG contends, however, that many of the deaths are the unintended result of drug abuse. The petition appears to suggest that in a search for euphoria, or because of a dependence on the drug, a user may take an excessive dose of propoxyphene or combine the drug with alcohol, narcotics, tranquilizers, sedative-hypnotics, or other substances that depress the central nervous system. The result can be an accidental death.

It is true that most identified propoxyphene-associated deaths appear to be the result of misuse of the drug, either in attempting suicide or in a drug abuse accident. In the report by Baselt et. al. (ref. 1), some of the cases classed as "accidental" involved such large quantities of propoxyphene that it is very likely that the drug was not being used for therapeutic purposes at the recommended dosage level.

Since filing the HRG petition, Dr. Wolfe has raised the question whether many of the deaths attributed to propoxyphene are due to a cariotoxic effect of its major metabolite, norpropoxyphene. This hypothesis, which would imply the existence of previously unidentified cases of propoxyphene-caused deaths possibly occurring at therapeutic doses of the drug, deserves serious consideration during FDA's review of the drug. At present, however, there is little evidence that this mechanism is a common factor in the deaths associated with propoxyphene.

Indeed, there is no clear evidence to date demonstrating that the therapeutic use of propoxyphene, in the absence of tranquilizers or alcohol, has caused accidental death. For example, although about one-third of the prescriptions for products containing propoxyphene are written for patients over age 60, these same patients experience only 8% of the deaths reported to be associated with propoxyphene. The largest incidence of deaths associated with propoxyphene products occurs among those in the 20-40 age range, who only receive about one

third of the prescriptions, but experience roughly half the deaths. If propoxyphene-associated deaths were predominantly accidental, one would expect a much higher proportion of the deaths to occur among users over 60, assuming that older users are at least as likely to have fatal accidents as younger users.

The only serious health risk from propoxyphene other than the deaths described above is that the drug can cause physical dependence. Otherwise, it does not cause significant adverse reactions in many cases. Miller and Greenblatt (ref. 3 found that adverse reactions in hospitalized patients are infrequent and mild. Moreover, although the adverse reactions from propoxyphene that did occur were qualitatively similar to those from codeine and other analgesics used in the hospital setting, they occurred less frequently. Standard tolerance studies in volunteers revealed no significant difference between propoxyphene and placebo. In contrast, Goodman and Gilman (ref. 4 state that in equianalgesic doses, propoxyphene and codeine may be expected to produce similar incidences of side effects.

Thus, the principal harm posed by propoxyphene, and the basis of the HRG petition, are the deaths associated with the use of the drug in suicide attempts or accidental overdosing or interactions with other nervous system depressants in drug abuse situations.

B. Possible Harm From Immediate Suspension of Propoxyphene From the General Market

The principal harm from immediate suspension of a drug is the loss to patients of the benefit of its therapeutic effectiveness. Therefore, to assess the harm from suspension of propoxyphene, it is necessary to evaluate the available information concerning its effectiveness.

I recognize that the efficacy of analgesics is particularly difficult to assess. Pain is a subjective symptom. I am informed that although it can be quantitatively measured for purposes of clinical trials, the conduct of such trials is complicated by the fact that any analgesic will have a large placebo effect, typically in the range of 30-35% of the patients. In addition, many experts believe that in the case of prescription analgesics, such as propoxyphene, the placebo effect associated with the drug is increased by the facts that the drug is prescribed by a physician after consultation with the patient, that the capsules and tablets are colored rather than white, and that the drug is dispensed by a pharmacist. Moreover, the overwhelming majority of prescriptions for products containing propoxyphene are for compounds containing it in combination with another analgesic, such as aspirin or acetaminophen. These combinations are clearly effective because of these other analgesics, and propoxyphene may make an additional contribution to their efficacy.

The literature on the efficacy of propoxyphene itself is mixed. HRG gives major attention to a literature review conducted by Miller et al. in 1970 (ref. 5). Miller cited 9 of 18 placebo controlled trials in which propoxyphene was found to be more effective than the placebo. Miller concluded that "[p]ropoxyphene is no more effective than aspirin or codeine and may even be inferior to these analgesics. .. When aspirin does not provide adequate analgesia it is unlikely that propoxyphene will do so." HRG also cites three subse quent studies that found no significant difference between propoxyphene and placebo. On the other hand, a 1978 study by Sunshine et al. (ref. 6) found propoxyphene napsylate at 200 mg (twice the recommended dose) to be significantly better than placebo. The low st dose used (50 mg) was slightly better than a placebo. The usual dose (100 mg) was not tested. In a second review of the literature in 1977, Miller (ref. 7) reported that three studies showed that propoxyphene is no more effective than a placebo and that five studies showed that it is as effective as (but not more effective than) a standard analgesic.

For purposes of this preliminary assessment of propoxyphene's efficacy in reaching an imminent hazard determination. I conclude that propoxyphene has some, but limited, efficacy.

Thus, it is possible that there may be some risk to patients who do not adequately respond to (or, in relatively few cases canrot safely take) aspirin. acetaminophen, or other analgesics, and who would be deprived of propoxyphene. Moreover, propoxyphene does induce some degree of physical dependence, so that suddent unavailability could lead to withdrawal symptoms for some patients.

Other patients who depend particularly on propoxyphene for relief from pain may experience some suffering as the result of the abrupt removal of the drug from the market. For these people, the most likely substitute for propoxyphene is codeine, which is widely believed to be even more addictive than propoxyphene. If presented with the sudden disappearance of propoxyphene from the market, physicians would still be reluctant to prescribe codeine for more than intermittent use, and patients would be reluctant to take it.

C. Likelihood of Final Action to Withdraw the Drug from the General Market'

The Bureau of Drugs in FDA has responsibility for initiating a withdrawal proceeding (21 CFR 314.200), but has not proposed that the NDA's for propoxyphene be withdrawn. Possible grounds for withdrawal of these NDA's include (1) that evidence from clinical experience shows the drug to be unsafe, (2) that new evidence not available when the NDA's were approved, together with the original evidence supporting the approvals, demonstrates that the drug is no longer shown to be safe, and (3) that the new evidence, evaluated together with the evidence in the original NDA's, supports a finding that there is a lack of substantial evidence that the drug is effective. 21 U.S.C. 355(e) (1), (2), and (3).

The issues concerning the safety and effectiveness of propoxyphene are difficult and complex.

Although the drug is associated with a large number of deaths, many of these deaths appear to be related to misuse of the drug rather than its use in accordance with the labeling directions. It is not clear that many of these deathsthose related to suicide attempts-would be prevented if propoxyphene were immediately removed from the market.

In addition, the record currently does not contain sufficient evidence for me to make a finding of imminent hazard based on two as yet unresolved issues raised by HRG's petition:

(1) The extent to which propoxyphene is dangerous, if at all, when used in accordance with the labeling;

(2) The extent to which labeling restrictions are effective in controlling use of propoxyphene that may lead to death."

On the basis of the information with respect to propoxyphene available to me at this time, I cannot conclude whether or not one or more of the new drug applications is likely to be withdrawn. That determination cannot be made until the issues concerning the efficacy and safety of propoxyphene in light of all the data now available have been developed more fully.

D. Potenial Alternative Means To Prevent Hazard

During the period FDA is evaluating further the safety and efficacy of propoxyphene, three steps can be taken to protect the public health. I am concerned by the various dangers posed by propoxyphene; use in suicides, accidental deaths from the interaction of the drug with alcohol or other drugs that act on the nervous system, and dependence on the drug. Therefore, I am directing that these problems be addressed immediately without awaiting the final FDA decision on whether propoxyphene meets the statutory standards of safety and effectiveness. I believe that implementation of the following actions will reduce the hazards to the public health.

First, the Department will promptly evaluate HRG's proposal to transfer propoxyphene from Schedule IV to Schedule II of the Controlled Substances Act. If this transfer were made, the production of propoxyphene would be limited by government-determined quotas; all distribution of the drug would be on special order forms; and prescriptions for the drug would not be refillable and would have to be in writing (i.e., telephone prescriptions would be prohibited). The Assistant

2 Because final responsibility for deciding whether the new drug applications for propoxyphene should be withdrawn is delegated to the Commissioner of Food and Drugs. I have not asked Dr. Kennedy to comment on this matter, and he has reserved judgment unti' formal administrative procedures have developed a complete record for his review. In the phenformin case, the evidence did support a finding that phenformin was dangerous even if used in accordance with the labeling. In addition, the evidence showed that phenformin was being used widely outside of the indications set out in the labeling.

Secretary for Health, who has delegated authority to make drug scheduling rec ommendations on behalf of the Department, will make a recommendation to the Department of Justice on propoxyphene in the near future, after consideration by FDA and its Drug Abuse Advisory Committee.

Second, FDA will expeditiously prepare and distribute appropriate information for physicians, dentists, and pharmacists regarding the risks associated with the use of propoxyphene. This information will encourage physicians and dentists to reconsider the risks of and need for the drug in specific cases. It should also help deal with the problems of suicide and accidental deaths from drug interactions by making physicians and dentists more cautious in prescribing the drug for patients who may be suicidal or who may be using alcohol or other drugs affecting the central nervous system. This information will also encourage pharmacists, when dispensing propoxyphene, to put on the container warnings against taking the drug in combination with tranquilizers or alcohol. Third, FDA will promptly prepare and distribute appropriate information for the general public, in the form of a published article or otherwise, regarding the risks associated with the use of propoxyphene.

Although I believe these actions will help protect the public, I do not believe that the completion and evaluation of these actions are necessary before a decision on the suspension or withdrawal of the propoxyphene NDA's can be made. VI. Conclusion

At this time, I do not believe that there is sufficient evidence available showing that the continued use of propoxyphene constitutes so serious a threat to public health as to warrant the extraordinary action of summary suspension of general distribution of the drug, pending initiation and completion of the procedures to determine whether propoxyphene should be removed permanently from the general market. Based on the record currently before me, I am unable to declare propoxyphene an “imminent hazard."

The Act carefully balances the safeguards against improvident withdrawals of NDA's and the need to protect the public health from significant risks. The suspension power vested in the Secretary should be used sparingly, when it is likely that the drug will ultimately be withdrawn from the market and immediate action will prevent serious harm during the pendancy of the withdrawal proceedings. The issues in the case of propoxyphene are in significant doubt, and I am not prepared to predict their outcome at this time.

The fact that I am not granting the HRG petition at this time does not mean that further evidence cannot lead me to an opposite conclusion. If, in the course of FDA's further review of propoxyphene, new information is developed to show that propoxyphene meets the criteria for suspension, I will act promptly. Furthermore, the other steps that I have directed should reduce the risks that propoxyphene poses to the public health, while FDA holds its hearing to determine whether the drug should be removed from the market.

Dated February 15, 1979.

JOSEPH A. CALIFANO, Jr., Secretary of Health, Education, and Welfare.

[From the New York Times, Feb. 18, 1979]

A COMPANY AT WAR: How LILLY DEFENDED DARVON-MARSHALING FORCES IN "RED FLAG ALERT"

(By Peter T. Kilborn)

INDIANAPOLIS.-In 1957, scientists of Eli Lilly & Company here introduced a painkiller that was safer and less addictive than the morphine and codeine that most physicians were then prescribing. The generic name of the drug was propoxyphene hydrochloride, and the brand name. Darvon. In due course, it became the third most prescribed drug in the United States. Then, on Nov. 21. 1978, the media relations director at Lilly, Russell Durbin, received a call from an Associated Press reporter in Washington. What, he asked, had Lilly to say about a petition to ban Darvon?

Thus began a long and wrenching episode for the 103-year-old giant of the pharmaceutical industry. With insulin, the Salk antipolio vaccine in the mid50's, and an ever-growing stable of antibiotics, Lilly has prided itself on doing