Correspondence

Valvular Heart Disease Associated with Fenfluramine–Phentermine

N Engl J Med 1997; 337:1772-1776December 11, 1997

Article

To the Editor:

Nineteen of the 24 patients described by Connolly et al. (Aug. 28 issue)1 in the primary report linking fenfluramine and phentermine with cardiac valvular abnormalities were identified in Fargo, North Dakota, a city with only 74,111 residents in 1990.2 Twelve of the Fargo cases, and perhaps the association itself, were initially identified by an astute echocardiographic sonographer. This relation had not been discovered previously despite the fact that fenfluramine has been approved since 1973 and that 18 million prescriptions for these drugs were filled in 1996 alone.1 Diagnosis of a cardiac abnormality in 18 of the 24 patients was based on echocardiographic findings of valvular insufficiency and “unusual valvular morphology.” 1 The key unanswered question is whether these echocardiographic abnormalities have clinical import for the majority of patients using these drugs or whether they are simply a visual finding. The drugs are so widely used and these reports have prompted such widespread concern that all efforts should be taken to be certain that the association with important heart disease is real. On the other hand, if the incidence rate in Fargo is confirmed and if the echocardiographic abnormalities frequently lead to clinically significant disease, these drugs will be responsible for a major epidemic of valvular heart disease that could have been far worse if not detected by the echocardiographer.

Paul D. Thompson, M.D.
Hartford Hospital, Hartford, CT 06102

2 References

1. 1

   Connolly HM, Crary JL, McGoon MD, et al. Valvular heart disease associated with fenfluramine-phentermine. N Engl J Med 1997;337:581-588
   Full Text | Web of Science | Medline

2. 2

   1990 Census. Compton's interactive encyclopedia, 1996 ed. Version V.0.3.kgb.
   

To the Editor:

Connolly et al. report valvular heart disease in 24 women treated with fenfluramine–phentermine who had no history of cardiac disease. They conclude that candidates for fenfluramine–phentermine therapy should be informed about serious potential cardiac adverse effects. We would like to extend this recommendation to other substances and other combinations of anorectic agents.

From October 1994 to July 1997, the Belgian Center for Pharmacovigilance received 43 reports of valvular heart disease among women who used anorectic drugs. Echocardiographic confirmation was available in 31 cases. Past exposure to anorectic drugs in these 31 women included treatment with fenfluramines (fenfluramine or dexfenfluramine) alone (8 women), fenfluramines and diethylpropion (18), fenfluramines and phentermine (1), and fenfluramines, diethylpropion, and phentermine (3); there was also 1 unconfirmed report of exposure to fenfluramines. Herbal medicine was also associated with 11 cases of exposure to fenfluramines and diethylpropion and 1 case of exposure to fenfluramines, phentermine, and diethylpropion. The mean (±SD) duration of exposure was estimated in 25 patients to be 39.6±28.1 months, and the mean age of the patients was 47.2±9.2 years.

Aortic-valve disorder was the only lesion seen in 18 patients. The other patients presented with lesions of the mitral valve alone (one patient), the aortic and mitral valves (seven), the aortic and tricuspid valves (one), and the aortic, mitral, and tricuspid valves (four). Thirteen patients presented with symptoms: dyspnea in six patients, primary pulmonary hypertension in two patients, and chest pain, chest discomfort, angina, pulmonary edema, and peripheral edema in one patient each.

There was exposure to fenfluramines in all cases, but our series differs from the cases described by Connolly et al. because of the high rate of exposure to the combination of fenfluramines and diethylpropion. In the absence of a clear physiopathologic explanation for the association between valvular disease and fenfluramine–phentermine, warnings issued regarding fenfluramine and phentermine should also apply to other combinations of anorectic agents. Moreover, caution is needed before one shifts a prescription from fenfluramine–phentermine to alternative drugs that are also associated with primary pulmonary hypertension.1

Xavier Kurz, M.D.
Ann Van Ermen, Pharm.D.
Belgian Center for Pharmacovigilance, 1010 Brussels, Belgium

1 References

1. 1

   Abenhaim L, Moride Y, Brenot F, et al. Appetite-suppressant drugs and the risk of primary pulmonary hypertension. N Engl J Med 1996;335:609-616
   Full Text | Web of Science | Medline

To the Editor:

Low-cost echocardiographic valvular screening for asymptomatic persons taking fenfluramine and phentermine was initiated on August 25, 1997, as a community service by Hendricks Community Hospital in Danville, Indiana, in association with the Hewlett–Packard Company. This screening was in response to the public health advisory, issued July 8, 1997, by the Food and Drug Administration (FDA), which described valvular heart disorders among patients treated with a combination of fenfluramine and phentermine. Table 1Table 1Echocardiographic Findings in 5 Men and 23 Women Treated with Fenfluramine and Phentermine. summarizes our initial findings for the first 5 men and 23 women who were screened.

Of the 14 female subjects with echocardiographically detected aortic insufficiency, the aortic insufficiency appeared to be pathologic in 10, 5 of whom had associated echocardiographically significant mitral regurgitation. Aortic insufficiency in the remaining four was minimal but abnormal for the subject's age.1 Only one woman was found to have isolated mitral regurgitation. All 28 subjects were unsolicited and all were clinically asymptomatic. Cardiac valves appeared to be echocardiographically normal in all 28 subjects, and there was no evidence of pulmonary hypertension. Color-flow Doppler echocardiography revealed abnormal valvular disorders in 16 of the 28 subjects, with a preponderance of the abnormalities detected in the female subjects.

On the basis of these early data, we urge echocardiographic surveillance of all persons being treated with fenfluramine–phentermine, with particular attention paid to asymptomatic women.

Susan Rasmussen, M.S.N.
Betty C. Corya, M.D.
Robert D. Glassman, M.D.
Hendricks Community Hospital, Danville, IN 46122

1 References

1. 1

   Klein AL, Burstow DJ, Tajik AJ, et al. Age-related prevalence of valvular regurgitation in normal subjects: a comprehensive color flow examination of 118 volunteers. J Am Soc Echocardiogr 1990;3:54-63
   Medline

To the Editor:

We read with alarm the article by Connolly and coauthors about the association between appetite suppressants and valvular heart disease. At that time, we were in the midst of a double-blind, placebo-controlled trial of fenfluramine and phentermine as treatment for type II diabetes mellitus. Connolly et al. described valvular damage with histopathological features identical to those seen in carcinoid heart disease and hypothesized that “the combination of fenfluramine and phentermine may potentiate the effect or concentration of circulating serotonin.” We started measuring serum serotonin and urinary 5-hydroxyindoleacetic acid in our study subjects. The serum serotonin determinations were performed by Associated University Regional Pathologists using high-pressure liquid chromatography. The normal range was 90 to 210 ng per milliliter.

On September 15, 1997, when fenfluramine and dexfenfluramine were withdrawn from the U.S. market, we unmasked our study and stopped treatment with fenfluramine and phentermine. Before doing so, we were able to measure serum serotonin levels during fenfluramine and phentermine treatment in seven subjects. The mean duration of treatment was 11 months (range, 2 to 16). To our surprise, none of these subjects had an elevated serum serotonin level, and five of the seven actually had serum serotonin levels that were below normal. Urinary 5-hydroxyindoleacetic acid values were normal in the three subjects in whom it was measured. The mean serum serotonin value in this group was 75 ng per milliliter (range, 34 to 134), which was somewhat lower than the corresponding mean value of 94 ng per milliliter (range, 48 to 138) in five placebo-treated subjects.

In patients with valvular heart damage due to the carcinoid syndrome, the mean serum serotonin level was 1720 ng per milliliter, a level that was more than six times the upper limit of normal, which was 264 ng per milliliter.1 Since the serum serotonin levels in our seven patients were far below the levels reported in carcinoid heart disease, we believe it is unlikely that valvular heart damage related to fenfluramine–phentermine is due to elevated circulating levels of serotonin. Consistent with this belief are reports that fenfluramine reduces whole-blood serotonin levels in children with autism2 and a report of reduced blood serotonin and urinary 5-hydroxyindoleacetic acid excretion in a woman with the carcinoid syndrome.3 We wonder whether the drugs or their metabolites have a direct action on serotonin receptors.

Bruce Redmon, M.D.
Susan Raatz, R.D., Ph.D.
John P. Bantle, M.D.
University of Minnesota, Minneapolis, MN 55455

3 References

1. 1

   Robiolio PA, Rigolin VH, Wilson JS, et al. Carcinoid heart disease: correlation of high serotonin levels with valvular abnormalities detected by cardiac catheterization and echocardiography. Circulation 1995;92:790-795
   Web of Science | Medline

2. 2

   Ho HH, Lockitch G, Eaves L, Jacobson B. Blood serotonin concentrations and fenfluramine therapy in autistic children. J Pediatr 1986;108:465-469
   CrossRef | Web of Science | Medline

3. 3

   Stahl SM, Levin B, Freedman DX. Serotonin depletion by fenfluramine in the carcinoid syndrome. N Engl J Med 1982;306:429-429
   Web of Science | Medline

To the Editor:

In the article by Connolly et al., the accompanying editorial,1 and the related correspondence (Aug. 28 issue),2 there is no mention of the results of preclinical toxicity tests in animals, which in the past have always been required by the FDA. In such tests there is a likelihood of finding similar heart lesions in rodents (mice and rats) or other mammals (e.g., dogs), and one is left to wonder whether these were required in this instance or performed by the pharmaceutical companies concerned. The absence of any mention of such tests by clinical investigators as well as by the FDA suggests a sea change since the passage of the 1962 amendments to the Pure Food and Drug Law.

Frederick W. Wolff, M.B., B.S., M.D.
George Washington University School of Medicine, Washington, DC 20037

2 References

1. 1

   Curfman GD. Diet pills redux. N Engl J Med 1997;337:629-630
   Full Text | Web of Science | Medline

2. 2

   Cannistra LB, Davis SM, Bauman AG. Valvular heart disease associated with dexfenfluramine. N Engl J Med 1997;337:636-636
   Full Text | Web of Science | Medline

To the Editor:

Dr. Curfman's editorial is inappropriately alarmist in tone, and its conclusions overreach the data. In fact, some of the data are either misinterpreted or misrepresented. Alarms (such as a recent oral-contraceptive-pill scare in the United Kingdom1) based on incomplete or insufficient information prevent clinicians or patients from making informed, evidence-based choices. Nevertheless, I do agree that the recent findings of valvular heart disease among users of the combination of fenfluramine and phentermine have warranted an alert, and in that regard, I applaud the Journal's waiver of the Ingelfinger Rule and the prudent stance adopted by the FDA to date.

The report by Connolly et al. of a case series is important, but it is uncontrolled and highly vulnerable to referral bias at the Mayo Clinic.2 The study is valuable for the generation of hypotheses as well as representing a responsible alert to clinicians. The need for two hypothesis-testing case–control studies is urgent and compelling. One study should replicate the European study on primary pulmonary hypertension in North America. The second one should define affected patients as those with valvular heart disease. The studies should be done with the collaboration of seasoned epidemiologists and experienced clinician-specialists.

As pointed out by Curfman, the biologic mechanism of interactions in the observed associations between appetite suppressants and pulmonary or cardiac pathology are unknown. These need to be pursued in further research — fundamental, clinical, and epidemiologic. In particular the “chicken and egg” conundrum is important to elucidate: Does rapid loss of weight cause the observed findings, or are any of the appetite suppressants themselves the cause of the adverse outcomes? An analogous puzzle is still unresolved. It concerns paradoxical early worsening of retinopathy in patients with diabetes when blood sugar is brought down precipitously with good control.3 Associations that emerge in observational and uncontrolled studies are not always causal.

As an epidemiologist I am confident that there will be no epidemics of primary pulmonary hypertension or cardiac valvular disease. The base-line incidence and prevalence of the disorders are so low that even statistically significant high odds ratios are extremely unlikely to result in population-defined incidence rates deemed to be of public health or clinical importance. On the other hand, the ongoing epidemic of obesity in America is real and remains a major one with respect to common killing and disabling cardiovascular consequences. The use of the word “outbreaks” in the editorial is unfortunate. Clusters of any uncommon disorder in highly specialized institutions such as the Mayo Clinic and Antoine Beclere Hospital in Paris mean little until referral bias, very strong in such prestigious centers, is taken into account.

It is regrettable that an editorial on scientific issues should portray a drug as guilty by association through the use of its trade name as a double-entendre in the title, “Diet Pills Redux.” This is especially inappropriate at a time when the evidence is not all in. Dexfenfluramine (Redux) alone has not been shown to be associated with any of the outcomes of concern. The publicity bias fueled by this tactic and the selective recall caused by the publicity will only make the necessary and complicated epidemiologic research yet to be done even more difficult.

Walter O. Spitzer, M.D.
Stanford University, Palo Alto, CA 94309-8608

3 References

1. 1

   Spitzer WO. Rebuilding confidence in oral contraceptives: a new imperative in family planning. Br J Clin Pharmacol 1996;41:359-363
   CrossRef | Web of Science | Medline

2. 2

   Connolly HM, Crary JL, McGoon MD, et al. Valvular heart disease associated with fenfluramine-phentermine. N Engl J Med 1997;337:581-588
   Full Text | Web of Science | Medline

3. 3

   The Diabetes Control and Complications Trial Research Group. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med 1993;329:977-986
   Full Text | Web of Science | Medline

To the Editor:

A disturbing feature of Curfman's editorial was his nihilistic approach to the treatment of obesity. He seems to assume that all physicians who prescribe anorectic agents do so in a vacuum, making no effort to train patients to become better lifetime weight managers. Curfman is correct, however, in his reference to the unlikelihood that medication alone will be a long-term solution for weight control. In any event, nihilism has never cured, controlled, or prevented any ailment.

Curfman says that “we should demand a moratorium on the use of anorectic drugs for the purpose of cosmetic weight loss.” As any physician specializing in bariatrics knows, even patients who are 50 percent or more above their ideal body weight often give only cosmetic reasons for wanting to lose weight. On the other hand, patients just 10 to 15 lb overweight sometimes say that their family history of diabetes or cardiovascular disease is their principal concern. I hope that physicians and their properly informed, consenting patients will continue to be permitted to make their own choices regarding treatment of excess weight.

Edward Morris Marshall, M.D.
515 N. Sepulveda, Manhattan Beach, CA 90266

Author/Editor Response

The authors reply:

To the Editor: Although limited by their small number of patients, the report by Redmon et al. reemphasizes that the mechanism of valve injury in patients treated with appetite suppressants remains uncertain. We wonder whether any of the patients had echocardiographic evidence of valvular heart disease. If not, the possibility remains that fenfluramine–phentermine may elevate circulating levels of serotonin in some patients (though not in any of Dr. Redmon's cohort), which in turn promotes valvular abnormalities. Until correlative data are available in sufficiently large populations, all hypotheses, including the possibility of a direct action on receptors, remain speculative.

Dr. Spitzer has misunderstood the means by which the patients were accrued. The majority were never seen at the Mayo Clinic but were evaluated at MeritCare in Fargo, North Dakota, and thus are not subject to referral bias to the Mayo Clinic. Indeed, one of the compelling aspects of our report was the serendipitous simultaneous observations of similar lesions at two unaffiliated institutions. We agree that further evaluation of the frequency with which pulmonary hypertension and valvular heart disease are associated with appetite suppressants is warranted in North America, and such studies are planned, although the process may be complicated by the withdrawal of two of the diet medications. The confident comment about an epidemic of valvular heart disease being unlikely may be attributed to the fact that Spitzer's letter antedates the withdrawal of fenfluramine and dexfenfluramine on September 15, 1997. Concern about an epidemic of valvular heart disease is justifiable if, as suggested by current FDA data, in up to 30 percent of patients who have taken diet drugs even mild valve disease develops. The public health importance of this problem has not yet been established, and the clinical significance cannot be evaluated at this time. Preliminary evidence suggests it is not trivial. Among over 100 clinical and surgical cases of valvular heart disease associated with diet drugs in the United States that have been reported to the FDA in the past five months, there was at least 1 case of endocarditis and 4 deaths related to valve surgery. Although obesity may be an important health problem in the United States, the impact of appetite suppressants on the health consequences of obesity has not been definitively characterized. Thus, the health benefits of the medications may not justify the potential health costs. Unfortunately, both factors in this equation remain nebulous.

Dr. Thompson asserts that the relation between fenfluramine and valvular heart disease was not discovered previously despite long clinical use. In view of the number of cases brought to the attention of the FDA since our report and the growing awareness of the cursory methods of clinical evaluation among at least some practitioners, it seems likely that such an association (or any subtle or sporadic untoward consequences of treatment) was simply not sought or noticed. The letter from Drs. Kurz and Van Ermen reveals that, indeed, at least some preliminary concern regarding a possible relation had been aroused before our report. Moreover, their observations suggest that merely switching from one culprit combination of medication to other drug combinations or herbal remedies may not eliminate concern about safety.

The data of Rasmussen and colleagues are representative of accumulating evidence that there is a credible association between appetite suppressants and cardiac valvular abnormalities. However, the report is open to criticism (as, in some respects, is ours): data were not collected in a blinded fashion, the patients did not undergo pretreatment echocardiographic evaluation, findings on physical examination and the severity of regurgitation were not reported, and no control group of obese and nonobese subjects was examined. Until such information is available, together with data about incidence, natural history, and pathophysiologic mechanisms, our understanding will remain incomplete.

Heidi M. Connolly, M.D.
Michael D. McGoon, M.D.
Mayo Clinic, Rochester, MN 55905

Author/Editor Response

Dr. Curfman replies:

On September 15, 1997, just two weeks after my editorial appeared in the Journal, fenfluramine and dexfenfluramine were withdrawn from the market at the request of the FDA. The decision was prompted by an analysis of new echocardiographic data suggesting that the incidence of valvular heart disease associated with the use of fenfluramine–phentermine may exceed 30 percent. Case–control studies are in progress to determine the incidence more precisely. If this estimate is confirmed, it means that thousands of people may have been affected by valvular heart disease induced by these diet pills. I am not an epidemiologist, but with all due respect to Dr. Spitzer, I believe that this qualifies as an outbreak.

In the light of these developments, I wonder whether Dr. Spitzer still feels that my editorial was “inappropriately alarmist” and whether Dr. Marshall still feels that it was “nihilistic.”

Gregory D. Curfman, M.D.

Citing Articles (13)

Citing Articles

1. 1

   Agnès Fournier, Mahmoud Zureik. (2012) Estimate of deaths due to valvular insufficiency attributable to the use of benfluorex in France. Pharmacoepidemiology and Drug Safetyn/a-n/a
   CrossRef

2. 2

   L D Whigham, N V Dhurandhar, P S Rahko, R L Atkinson. (2006) Comparison of combinations of drugs for treatment of obesity: body weight and echocardiographic status. International Journal of Obesity
   CrossRef

3. 3

   Lisa L Ioannides-Demos, Joseph Proietto, Andrew M Tonkin, John J McNeil. (2006) Safety of Drug Therapies Used for Weight Loss and Treatment of Obesity. Drug Safety 29:4, 277-302
   CrossRef

4. 4

   2006. Fenfluramines. , 1333-1344.
   CrossRef

5. 5

   Richard B Rothman, Michael H Baumann. (2002) Therapeutic and adverse actions of serotonin transporter substrates. Pharmacology & Therapeutics 95:1, 73-88
   CrossRef

6. 6

   Richard B Rothman, Michael H Baumann. (2002) Serotonin releasing agents. Pharmacology Biochemistry and Behavior 71:4, 825-836
   CrossRef

7. 7

   Richard B. Rothman, Michael H. Baumann. (2000) Neurochemical mechanisms of phentermine and fenfluramine: Therapeutic and adverse effects. Drug Development Research 51:2, 52-65
   CrossRef

8. 8

   Neil J Weissman, John F Tighe, John S Gottdiener, John T Gwynne. (1999) Prevalence of valvular-regurgitation associated with dexfenfluramine three to five months after discontinuation of treatment. Journal of the American College of Cardiology 34:7, 2088-2095
   CrossRef

9. 9

   Richard W. Asinger. (1999) The Fen-Phen Controversy: Is Regression Another Piece of the Puzzle?. Mayo Clinic Proceedings 74:12, 1302-1304
   CrossRef

10. 10

    Richard B. Rothman. (1999) Is phentermine an inhibitor of monoamine oxidase? A critical appraisal. Synapse 32:2, 141-145
    CrossRef

11. 11

    Laura E. Derby, Marian W. Myers, Hershel Jick. (1999) Use of dexfenfluramine, fenfluramine and phentermine and the risk of stroke. British Journal of Clinical Pharmacology 47:5, 565-569
    CrossRef

12. 12

    Weissman, Neil J., Tighe, John F. Jr., Gottdiener, John S., Gwynne, John T., . (1998) An Assessment of Heart-Valve Abnormalities in Obese Patients Taking Dexfenfluramine, Sustained-Release Dexfenfluramine, or Placebo. New England Journal of Medicine 339:11, 725-732
    Full Text

13. 13

    Charles Van Ypersele de Strihou. (1998) Valvular heart disease and Chinese-herb nephropathy. The Lancet 351:9107, 991
    CrossRef