Fenfluramines

Cardiomyopathy

Restrictive cardiomyopathy due to endocardial fibrosis occurred in a 35-year-old woman 5 months after she had started to take fenfluramine 10 mg tds and phentermine 15 mg/day [ ]. The endocardial findings strongly resembled the valvular lesions associated with the use of fenfluramine–phentermine. Endocardial and valvular fibrosis associated with anorectic drugs is strikingly similar to the plaque material found in patients with carcinoid syndrome and those exposed to methysergide, and all possibly arise from a common mechanism.

Valvulopathy

The fenfluramines and phentermine can cause valvular heart disease [ ], and this has been reviewed [ ]. Fenfluramine was voluntarily withdrawn by the manufacturers on 15 September 1997, and the US Department of Health and Human Services issued interim recommendations for people previously exposed to fenfluramine or dexfenfluramine with cardiac valvulopathies [ ].

The use of fenfluramine or dexfenfluramine alone or in combination with phentermine, in 2524 adult participants in the population-based Hypertension Genetic Epidemiology Network Study, was associated with aortic regurgitation independent of aortic dilatation or fibrocalcification [ ]. The association between the use of fenfluramine or dexfenfluramine (alone or with phentermine) and aortic regurgitation adjusted for potential confounders was analysed. Nineteen participants, all of whom had hypertension, were being treated with fenfluramine or dexfenfluramine (5 on these agents alone, 14 also with phentermine). Aortic regurgitation was present in 32% (n = 6) of those taking fenfluramine/dexfenfluramine versus 6% (162/2505) of the remaining subjects. In multivariate analyses, after adjusting for important confounders, in particular aortic root structure, treatment with fenfluramine or dexfenfluramine was associated with aortic regurgitation (OR, 5.2; 95% CI, 1.7–14) and fibrocalcification (OR, 5.2; 95% CI, 1.9–15).

The autopsy findings in the heart and lungs of a patient with pulmonary hypertension associated with fenfluramine and phentermine have been described [ ].

• A 36-year-old woman with a body mass index of 47.5 kg/m ² , took fen–phen for 7 months and developed pulmonary hypertension. Her pulmonary arterial pressure was 56 mmHg and echocardiography showed right ventricular dilatation and hypokinesia. She had a cardiopulmonary arrest during right-heart catheterization and died 3 days later. At autopsy, there was right ventricular dilatation with a fibroproliferative tricuspid valve. The pulmonary arteries had fibroproliferative plaques which were more severe and prominent in the upper lobes than in the lower lobes.

More autopsy cases of patients with a history of fen–phen use are warranted to document the frequency of combined cardiac valvular disease and pulmonary hypertension.

Progressive pulmonary hypertension occurred in two patients who took fenfluramine for only 8 months [ ]. The symptoms abated on withdrawal but returned in one patient when rechallenged.

In 1996, in a case-control study, 95 patients from 35 centers in France, Belgium, the UK, and the Netherlands were compared with 355 age- and sex-matched controls [ ]. The use of anorexic drugs (mainly derivatives of fenfluramine) was associated with an increased risk of primary pulmonary hypertension. Association with recognized risk factors such as a family history of primary pulmonary hypertension, infection with HIV, or the use of intravenous drugs was also confirmed. The absolute risk for obese patients who took anorexic agents for more than 3 months was 30 times higher than in non-users.

Seven years after fenfluramines were withdrawn, reports of mitral and aortic valvular disease continue to appear [ ].

• When a 39-year-old woman developed pneumonia her systolic and diastolic murmurs, which had been previously detected, were re-evaluated. Echocardiography showed grade III mitral regurgitation and grade II/III aortic regurgitation. The valves were pearly white and retractile, with no commissural symphysis or calcification. From the age of 19 years she had taken intermittent treatment with anorectics for 10 years (cumulative total of 34 months of fenfluramine 60 mg/day and 3 months of dexfenfluramine 45 mg/day). The murmurs were first detected at the age of 25 years, 6 years after she had started taking anorectics. Histological examination of the resected valves showed proliferation of myofibroblasts expressing smooth muscle α-actin.

This case suggests that severe valvular regurgitation associated with fenfluramines can occur late. Long-term health surveillance of anorectic exposed subjects may be necessary.

The evidence implicating ion channels in pulmonary artery smooth muscle vasoconstriction, proliferation, and/or reduced apoptosis has been reviewed [ ]. Pulmonary arterial hypertension may be considered as a form of vascular channelopathy. Experiments performed on cloned, voltage-activated potassium channels (K _v channels) in expression systems have confirmed that dexfenfluramine inhibits cloned K _v2.1 channels expressed in Xenopus laevis oocytes [ ]; a number of anorectic drugs, including aminorex, dexfenfluramine, phentermine, and sibutramine, inhibit K _v1.5 channels expressed in CHO cells [ ]. To reduce the risk of future epidemics of pulmonary arterial hypertension, it has been suggested that potential appetite suppressant molecules should be screened in heterologous K _v channel expression systems.

Echocardiography with color Doppler in 22 patients aged 25–69 years (19 women and three men) who had taken fen–phen for more than 3 months showed that one patient with newly discovered aortic insufficiency was asymptomatic. Some were taking several other drugs, none of which is known to precipitate valvular heart disease. Echocardiography was normal in 12 cases and abnormal in 10 including significant aortic insufficiency and significant mitral regurgitation. Ten of the patients had significant aortic insufficiency and nine had at least mild mitral insufficiency. The author inferred that fenfluramine was the likely offending agent, because (a) while it is known to cause release of serotonin, phentermine does not; (b) carcinoid tumors, which secrete serotonin and ergotamine, a serotonergic drug, are known to cause valvular heart disease; and (c) none of the obese patients who took phentermine and fluoxetine for more than 2 years developed pulmonary hypertension [ ] and the author found no valvular heart disease in this cohort either [ ]. The recommendation that phentermine should be combined with fluoxetine, sertraline, or fluvoxamine as safer alternatives [ ] requires prospective studies.

It was suggested that in patients who met the FDA criteria for cardiac valvular abnormalities on echocardiography performed soon after the withdrawal of appetite suppressants, there was a possibility (ranging from as low as 5% to as high as 67%) that the abnormality was a naturally occurring phenomenon and not a consequence of drug use [ ]. However, various studies have supported earlier reports of an association between fenfluramine or its d-isomer and cardiac valvular regurgitation, although they have differed with regard to the strength and clinical significance of the association [ ]. Differences in design, including a lack of baseline cardiac evaluation in echocardiographic assessment [ , , ], have precluded comparisons. Additional evidence linking the use of fenfluramine or dexfenfluramine to cardiac valvular regurgitation has reaffirmed the wisdom of the FDA’s decision to withdraw them from the market.

Why was this type of valvulopathy not recognized sooner? Changes in medical practice seem to have played a role as long-term and widespread use of these drugs evolved in the 1990s. Furthermore, cardiac murmurs can be more difficult to detect in obese patients [ ].

Prevalence . Although initially a prevalence of up to 30% was estimated, subsequent reports have suggested much lower rates. There are several reasons for this disparity including uncontrolled data and the limitations of echocardiography [ ]. The FDA surveys and the University of Minnesota study reported point prevalences and inherently overestimated the association of appetite suppressants with valvulopathy because a certain percentage of patients have pre-existing valvular lesions. The method of detection also plays a crucial role. Echocardiography is far more sensitive than clinical examination in detecting valvular regurgitation. The issue may also be confounded by lesion regression after withdrawal of therapy [ ]. Moreover, case-control studies are no substitute for objective evidence of the status of cardiac valves before drug exposure [ ]. Also the duration of exposure has varied widely in different reports.

Pulmonary hypertension and valvular heart disease associated with anorexigens have been described predominantly in women, which raises important questions about biological and psychological risk factors and ethical practice. Do women respond differently to these drugs because of genetic or physiological factors or are these drugs being prescribed almost exclusively for women? Was it realistic for the regulatory authorities to believe that these drugs would be used only to treat morbid obesity? Most important, what view of the benefit to harm balance of using anorexigenic drugs has allowed women and their physicians to justify the use of potentially lethal drugs to deal with concerns about body image and weight? These questions [ ] are pertinent to the current scenario of appetite-suppressant drug-related concerns.

Studies in which baseline echocardiography was carried out before the drug was used showed that the risk of new or prospective valvular heart disease was much lower than implied by previous prevalence studies [ ]. In 46 patients who used fenfluramine or dexfenfluramine for 14 days or more, the primary outcome was new or worsening valvulopathy, defined as progression of either aortic regurgitation or mitral regurgitation by at least one degree of severity and disease that met FDA criteria. Two patients taking fen–phen developed valvular heart disease. One had mild aortic regurgitation that progressed to moderate regurgitation and the second developed new moderate aortic insufficiency. The authors argued that the referral bias in their study, which required an echocardiogram for inclusion, would have tended to result in a higher incidence of valvular disease.

In an amended randomized double-blind placebo-controlled comparison of dexfenfluramine with an investigational modified-release formulation of dexfenfluramine, the study medication was discontinued and echocardiographic examinations were performed on 1072 overweight patients within a median of 1 month after withdrawal of treatment [ ]. These patients, 80% of whom were women, had been randomly assigned to receive dexfenfluramine (366 patients), modified-release dexfenfluramine (352 patients), or placebo (354 patients). The average duration of treatment was 71–72 days in each group. Echocardiograms were assessed blind. Pooling the fenfluramine groups, there was a higher prevalence of any degree of aortic regurgitation (17% versus 12%) and mitral regurgitation (61% versus 54%) with fenfluramine. Analyses carried out using the criteria set by the FDA showed that aortic regurgitation of mild or greater severity occurred in 5% of the patients taking dexfenfluramine, 5.4% of those in the two fenfluramine groups combined, and 3.6% of those in the placebo group. Moderate or severe mitral regurgitation occurred in 1.7% and 1.8% of those taking fenfluramine and 1.2% of those taking placebo. Aortic regurgitation of mild or greater severity, mitral regurgitation of moderate or greater severity, or both occurred in 6.5%, 6.9%, and 4.5% respectively.

This was an unusual study because patients enrolled for a different purpose were analysed mid-way through the study in response to withdrawal of fenfluramine. Exposure to fenfluramine in this study was relatively short (2–3 months) and the prevalences of mitral regurgitation and aortic regurgitation in this study were much lower than previously described [ ].

Although the findings of this study may be reassuring for patients who have taken dexfenfluramine for 2–3 months, they should not preclude the appropriate investigation of a new murmur or new symptoms in any patient with a history of exposure to dexfenfluramine as specified in the American College of Cardiology Guidelines [ ].

In 24 women who were evaluated an average of 12 months after starting to take fen–phen, echocardiography showed that all had unusual valvular morphology and regurgitation affecting valves on the right and left sides [ ]. Eight women also had newly documented pulmonary hypertension. Histopathological findings included plaque-like encasement of the leaflets and chordal structures with intact valve architecture. The histopathological features were like those seen in carcinoid or ergotamine-induced valve disease. As of the end of September 1997, the FDA had received 144 individual spontaneous reports (including the 24 cases reported earlier) involving fenfluramine or dexfenfluramine with or without phentermine in association with valvulopathy, 113 with complete information [ ]. Of these, 98% occurred in women of whom 2% used fenfluramine alone, 14% used dexfenfluramine alone, 79% used fenfluramine with phentermine, and 5% used a combination of all three; none had used phentermine alone. The median duration of drug use was 9 months (range 1–39 months). Cardiac valve replacement surgery was required in 24% and there was an 11% mortality.

Based on a prospective study carried out in 226 obese subjects (183 women and 43 men) with a mean body mass index of 40 kg/m ² , therapy with fen–phen was associated with low prevalence of significant valvular regurgitation [ ]. The authors suggested that valvular regurgitation in these subjects may have reflected age-related degenerative changes. However, several limitations of this study have been pointed out: (a) there was no control group and not all the subjects had echocardiography; (b) multiple readers interpreted the echocardiograms, rendering the comparison less accurate; (c) there was inherent inaccuracy in differentiating mild degrees of valvular regurgitation especially using qualitative scoring systems; (d) there was selection bias; and (e) neither direct inspection nor histopathological confirmation of valvular lesions was performed on any patient.

The risk of a subsequent clinical diagnosis of a valvular disorder of uncertain origin has been assessed in a population-based follow-up study using nested case-control analysis of 6532 subjects who took dexfenfluramine, 2371 who took fenfluramine, and 862 who took phentermine [ ]. The control group comprised 9281 obese subjects who did not take appetite suppressants matched with the treated subjects for age, sex, and weight. No subject had cardiovascular disease at the start of the follow-up for an average duration of 5 years. There were 11 cases of newly diagnosed idiopathic valvular disorders, five with dexfenfluramine and six with fenfluramine. There were six cases of aortic regurgitation, two of mitral regurgitation, and three of combined aortic and mitral regurgitation. There were no cases of idiopathic cardiac valve abnormalities among the controls or those who took phentermine. The 5-year cumulative incidence of idiopathic cardiac valve disorders was 0 per 10 000 among both those who had not taken appetite suppressants (95 CI = 0, 15) and those who took phentermine alone (CI = 0, 77), 7.1 per 10 000 among those who took either fenfluramine or dexfenfluramine for less than 4 months (CI = 3.6, 18), and 35 per 10 000 among those who took either of these medications for 4 months or more (CI = 16, 76). The authors concluded that the use of fenfluramine or dexfenfluramine, particularly when used for 4 months or longer, is associated with an increased risk of newly diagnosed cardiac valve disorders, particularly aortic regurgitation.

The above study was based on information derived from the General Practice Research Database in the UK. Subjects who had been given at least one prescription for dexfenfluramine, fenfluramine, or phentermine after 1 January 1988, and who were 70 years or younger at the time of their first prescription were included. Subjects were considered to have a new cardiac abnormality if they had no history, on the basis of clinical records, of cardiac valvular abnormalities and if there was evidence of a new valvular disorder on the basis of echocardiography or clinical examination after exposure to appetite suppressants. All the data had been recorded before the publication of recent reports of an association between appetite suppressants and primary pulmonary hypertension [ ] or cardiac valve disorders [ ], reports of which continue to appear [ ]. Hence, it was possible to exclude the possibility that enhanced awareness of possible serious adverse reactions to appetite suppressants had led to closer surveillance of patients who were taking these drugs. Nevertheless, the study did not provide information on the frequency of idiopathic cardiac valve disorders that are asymptomatic or otherwise not clinically diagnosed.

Using the FDA case definition of appetite-suppressant related valvulopathy, the prevalence was 31% (60/191) in a selected group of Mayo Clinic patients at Rochester [ ]. The most common finding was mild aortic regurgitation. Of asymptomatic patients 28% had abnormal echocardiographic findings. This study emphasized the spectrum of diet- or drug-related cardiac disease and the potential for valvulopathy in asymptomatic patients.

In patients who had taken dexfenfluramine (n = 479) or fen–phen (n = 455) continuously for 30 days or more in the previous 14 months, there was an increase in the prevalence of aortic regurgitation compared with 539 control subjects [ ]. There was no increase in the prevalence of moderate or severe aortic regurgitation in treated patients, and no difference in the prevalence of mitral regurgitation between the untreated and treated groups, irrespective of duration of therapy. All evaluations were carried out using the FDA criteria. The authors were careful to point out that their study was not specifically designed or adequately powered to evaluate specific categories of anorexigen therapy duration.

Further evidence that the prevalence of significant valvular regurgitation is low in patients who take fen–phen has been reported [ ]. Transthoracic echocardiography was performed in 343 obese patients in a 3-year prospective study that began within 4 months from the withdrawal of fenfluramine and dexfenfluramine from the market. There were 281 women and 62 men, mean age 47 years, and mean body mass index 40 kg/m ² . Using the FDA’s criteria, only 21 subjects (6.1%) had significant valvular lesions. Aortic regurgitation was detected in 18 subjects, mitral regurgitation in 3, and both aortic and mitral regurgitation in 1. Significant valvular disease did not correlate with age, sex, initial or final body mass index, drug dose, or the duration of therapy.

Mechanisms and susceptibility factors . The determinants of valvulopathy in patients treated with dexfenfluramine have been investigated: age and blood pressure can also affect the prevalence of regurgitation [ , ], as can duration of exposure [ ]. Others have found no correlation between valvular disease associated with appetite suppressants and either dose or duration of drug exposure [ ].

Cases of severe diffuse multivalvular disease associated with fen–phen have been described [ , ].

• A 52-year-old woman had a transesophageal echocardiogram 1 year before starting to take fen–phen, and had no significant valvular disease. She presented a year later with a new heart murmur and eventually required isolated aortic valve replacement. Pathological evaluation of the excised aortic valve was consistent with that described with fen–phen use.

• A 44-year-old woman who had previously taken appetite suppressants, developed valvular disease consistent with the effects of fen–phen. She had an identical twin who, despite having been treated with the same medication, remained symptom-free and without abnormal echocardiography. Both the patient and her sister took fen–phen for 2 years. However, the patient took a daily dose of fenfluramine of 60–120 mg (and often as much as 240 mg) and phentermine 90 mg (at times 180 mg), whereas her twin sister adhered to the daily amount prescribed (fenfluramine 60 mg and phentermine 24 mg).

The latter case suggests that dosage is important in the production of the valvular pathology. Mitral and tricuspid insufficiency developed in a 36-year-old woman who had taken fen–phen for 24 months [ ]. Transmission electron microscopy of the mitral and tricuspid valves showed many areas that appeared to contain intracellular, virus-like particles clustered in the cytoplasm, with a mean diameter of 32 nm. Whether this finding was incidental or related to the underlying pathology was uncertain.

There is further evidence, from an uncontrolled observational study in 85 patients, that the dose and duration of administration of fen–phen affects the risk of significant valvular disease [ ]. The authors suggested that it would be prudent to consider diagnostic echocardiography in patients who have used fen–phen either in a dosage of at least 60 mg/day or for at least 9 months. They also raised concerns that for patients with mild obesity, the prolonged use of larger cumulative amounts may lead to a higher risk of valve regurgitation.

There is further evidence of the relation between the duration of treatment with fen–phen and the prevalence of valvular abnormalities [ ]. In 1163 patients who had taken anorexigens within the previous 5 years and 672 control patients who had not, valvular abnormalities primarily involved those who had taken anorexigens for more than 6 months, and predominantly resulted in mild aortic regurgitation. The study had some noteworthy limitations: since fenfluramine has been withdrawn from use, a randomized trial was impossible; also the lack of baseline echocardiograms before treatment implies that one cannot be certain that the valvular regurgitation developed subsequent to drug treatment.

Diagnosis . It has been proposed that valvular disease can be attributable to appetite suppressants only if the following criteria are satisfied:

• the macroscopic and microscopic features are consistent with fenfluramine-related valvulopathy;

• clinical, echocardiographic, and intraoperative findings support the diagnosis;

• the history of drug exposure predates the development or exacerbation of valvular dysfunction [ ].

It is obvious that these criteria can be applicable only in cases in which cardiac valves are explanted and are available for histopathological studies.

The prevalence and diagnostic value of cardiac murmurs for valvular regurgitation has been determined in 223 patients taking dexfenfluramine for 6.9 months and 189 matched controls. Experienced physicians, non-cardiologists, who were unaware of the echocardiographic findings, took a history and performed cardiac auscultation. Based on their findings the authors recommended that cardiac auscultation should be the screening method of choice for detecting valvular regurgitation in users of anorexigens [ ]. In this study, the absence of cardiac murmurs predicted the absence of clinically important valvular regurgitation in 93% of dexfenfluramine users. These results support the recommendation by the American Heart Association and American College of Cardiology [ ] that asymptomatic users of anorexigens without a cardiac murmur do not warrant echocardiography. The data also suggest that in users of anorexigens with a 10% prevalence of valvular regurgitation and a 10–15% prevalence of cardiac murmurs, cardiac auscultation will prevent 85–90% of patients from undergoing unnecessary echocardiography. These implications may apply to all users of anorexigens because the prevalence of valvular regurgitation in recent large series is similar to that in this study [ ]. There are therefore large potential cost savings of cardiac auscultation, by preventing a large proportion of the more than 6 million Americans who are exposed to anorexigens from undergoing initial and follow-up echocardiography or from receiving empiric antibiotic prophylaxis for emergency procedures that preclude further cardiac evaluation.

Effects of withdrawal of therapy on valvulopathy . In a patient who was followed-up for 2 years after withdrawal, multivalvular regurgitation associated with fenfluramine and phentermine may have regressed [ ].

• A 44-year-old woman with morbid obesity but no history of cardiac disease developed atypical chest pain. Myocardial infarction was ruled out, and an echocardiogram showed normal chamber sizes and mildly reduced global systolic function. However, moderate to moderately severe aortic regurgitation, mild mitral regurgitation, and moderate tricuspid regurgitation were present. The estimated pulmonary artery pressure was slightly raised. Her only medications were fenfluramine 60 mg/day and phentermine 30 mg/day, which she had taken for the previous 50 weeks, during which time she had lost 40 kg. These drugs were withdrawn and 6 months later an echocardiogram showed improved left ventricular function and a reduction in the severity of all her valvular lesions with no clinically significant change in the estimated pulmonary artery pressure. An echocardiogram obtained 2 years after the initial study showed only trace aortic and tricuspid regurgitation without mitral regurgitation.

In this case, serial echocardiography over 2 years documented regression of multivalvular regurgitation, first discovered while the patient was taking fenfluramine and phentermine. The authors argued that although she was also given lisinopril, the marked degree of improvement in all the valvular lesions after withdrawal of the appetite suppressants was unlikely to be attributable to this alone [ ].

The small increase in prevalence of minor degrees of aortic regurgitation and mitral regurgitation in 941 patients treated with dexfenfluramine for 2–3 months was no longer present 3–5 months (median 137 days) after withdrawal [ ]. Echocardiograms were acquired using a standardized protocol and were assessed blindly.

In 50 patients with fenfluramine-associated valvular heart disease followed by serial echocardiography for 6–24 months after withdrawal of therapy [ ], in most cases valvular heart disease either did not change or improved at least by one grade. Mitral and aortic regurgitation improved in some patients, and tricuspid and pulmonic regurgitation improved in most patients after withdrawal. When improvement did occur, regression of regurgitation often involved multiple valves on both the left and right sides of the heart, rather than affecting one valve in isolation. Although most of the patients stabilized or improved, a few had worsening of valvular regurgitation despite withdrawal.

Comparable results were also reported in a larger series in another study [ ]. Sequential echocardiographic evaluation 1 year after withdrawal of dexfenfluramine showed a significant reduction in aortic regurgitation. There were no significant changes in mitral regurgitation or any other valvular variables. Although these results can be applied only to the population studied (predominantly middle-aged, obese, white women who took dexfenfluramine for 2–3 months), the implications are considerable. Because valvular regurgitation remained stable or improved in most of the patients, surgical referral for patients with severe regurgitation may be delayed. Improvement in valvular regurgitation often occurred within months after drug withdrawal. Watchful waiting with serial echocardiography, prophylaxis against endocarditis, and medical therapy may be a reasonable management strategy in patients with severe regurgitation, minimal symptoms, and no evidence of left ventricular dysfunction [ ].

However, reversibility may not occur in all cases [ ].

• Cardiac allograft transplantation was carried out from a 35-year-old hypertensive donor with prolonged exposure to fenfluramine and phentermine. There was non-specific mitral valve thickening, with trivial mitral regurgitation and poor approximation of the mitral valve leaflets, due to reduced posterior leaflet mobility. There was no evidence of any other valvular lesion. Examination of the donor heart during cardiac implantation showed three discontinuous lesions along the left atrial surface of the mitral valve annulus and another firm nodular lesion of the annular endocardium. Transplantation of the heart was uneventful and intraoperative transesophageal echocardiography, performed after weaning from cardiopulmonary bypass, showed trivial mitral regurgitation with excellent allograft contractility. The postoperative course was uneventful and the patient was discharged on the eighth postoperative day. Histological examination of the specimen showed a glistening appearance with proliferating myofibroblasts and associated fibrinous vegetations. There was no evidence of acute or chronic inflammation, and Gram staining did not show bacterial or fungal elements. A transthoracic echocardiogram 6 weeks after transplantation showed only trivial mitral regurgitation with improved mobility of the posterior leaflet. Hemodynamic data showed normal allograft function. However, after 6 months of follow-up, Doppler echocardiography showed worsening of mitral regurgitation to moderate severity, but no adverse effects of this hemodynamic load were noted and the patient remained stable.

Although conclusions based on single cases have limitations, they can often provide useful insights and act as catalysts for further studies. This report illustrates a few important features of cardiac valvulopathy associated with anorexigen use. There was no involvement of chordal apparatus and so the pathological changes within the valve leaflets and annulus represented the earliest site of an anorexigen-induced valvulopathy. This opportunity to observe a case of “early” valvulopathy visually and histopathologically offered insight into pathogenesis, and may have been helpful in staging the lesions temporally.

Of 120 patients who had follow-up echocardiography at least twice after stopping fen–phen, 99 met FDA criteria for valvulopathy [ ]. On second echocardiography, 57 of these 99 had no change in valvulopathy, 33 had improved, and 9 had deteriorated; nine patients no longer met FDA criteria for valvulopathy. The authors suggested that physicians must continue to be vigilant with patients who develop valvulopathy after taking fen–phen.

Obesity as a confounding factor in cardiac valvulopathy . It is not clear whether valvular insufficiency is related to the use of appetite suppressants or is simply a consequence of obesity. Obese patients who took dexfenfluramine alone, dexfenfluramine in combination with phentermine, or fenfluramine in combination with phentermine have been compared with a matched group of obese-control subjects who had not taken these medications [ ]. A total of 1.3% of the controls (3 of 233) and 23% of the patients (53 of 233) met the case definition for cardiac valve abnormalities (OR = 23). The odds ratios for such cardiac valve abnormalities were 13 with dexfenfluramine alone, 25 with dexfenfluramine and phentermine, and 26 with fenfluramine and phentermine. This study showed that the prevalence of valvular insufficiency is significantly higher among obese patients who have taken appetite suppressants than among subjects matched for age, sex, and body mass index who did not take such drugs. Since a higher percentage of patients than controls had trace aortic valve insufficiency, the authors questioned whether the case definition threshold for cardiac valve abnormalities in association with appetite suppressants set by the FDA and Centers for Disease Control and Prevention is perhaps too high. In this study the factors that predisposed patients to valvular insufficiency were (a) age at the start of therapy, (b) use of dexfenfluramine, (c) combination of dexfenfluramine with phentermine, and (d) combination of fenfluramine with phentermine. Hence, neither the clinical significance nor the natural history of this type of valvular disease has been defined.

Other epidemiological studies have ruled out the possibility that obesity itself causes a high prevalence of cardiac valvular regurgitation [ ].

Pulmonary hypertension

Pulmonary hypertension associated with fenfluramine was first reported in the early 1980s. A retrospective study further established the link [ ]. Subsequently a multicenter case-control study in 95 patients showed a high incidence of pulmonary hypertension in patients who had used fenfluramine or dexfenfluramine (14). Moreover, there was a strong suggestion of a dose–response effect, longer periods of use being associated with a progressive increase in the relative risk of pulmonary hypertension. In 1997, the first case of pulmonary hypertension in association with fen–phen was reported [ ]. Eight of the 24 patients with valvular disease had newly diagnosed pulmonary hypertension, although in most cases it was attributable to valvular abnormalities [ ]. It is not clear whether a combination of these agents poses a higher risk in predisposed individuals.

The results of a Belgian study in 35 patients with pulmonary hypertension and 85 matched controls have been published [ ]. The data were collected when there was no restriction on prescribing of appetite suppressants. Of the patients, 23 had previously taken appetite suppressants, mainly fenfluramines, compared with five controls. Moreover, the patients who had been exposed to appetite suppressants tended to be on an average more severely ill and to have a shorter median delay between the onset of symptoms and diagnosis.

Pulmonary artery pressure and cardiac valvular status were determined in a series of 156 mostly asymptomatic patients taking fenfluramine and phentermine [ ]. The anorexigen was withdrawn when abnormalities were noted. Pulmonary artery pressure was estimated and valvular examination was performed using Doppler echocardiography. There was borderline or mildly elevated pulmonary artery pressure in 21 patients and 31 patients had notable valvular abnormalities. It has therefore been established that asymptomatic patients may have significant echocardiographic abnormalities, representing early lesions.

• A 30-year-old woman who had taken dexfenfluramine for 7 months developed pulmonary hypertension and right heart failure during late pregnancy. She died of septicemia with multiorgan failure 4 days after a cesarean section [ ].

• Pulmonary hypertension and multivalvular damage after prolonged use of fenfluramine with phentermine have been reported in a 70-year-old Israeli woman [ ].

• Fatal pulmonary hypertension occurred in a 32-year-old man who had been taking phentermine in unknown doses for 4 months [ ].

Incidence . The epidemiological association of pulmonary hypertension with aminorex and dexfenfluramine, both with respect to the strength of the association (estimate of relative risk) and its impact on public health, has been investigated [ ]. Control rates of exposure were used to estimate population exposure prevalences. The estimated odds ratio for the association between pulmonary hypertension and any exposure to aminorex was 98 and for dexfenfluramine 3.7. The strong association between aminorex and pulmonary hypertension projected a fivefold increase in the incidence of pulmonary hypertension, and thus a very noticeable epidemic. In contrast, the association with dexfenfluramine is expected to result in an incidence of only 20% and thus a repeat epidemic seems unlikely.

In a prospective surveillance study of 579 patients with pulmonary hypertension at 12 large referral centers in North America, 205 had primary pulmonary hypertension and 374 had secondary pulmonary hypertension [ ]. Among the drugs surveyed, only fenfluramine had a significant association with primary pulmonary hypertension compared with secondary pulmonary hypertension (adjusted odds ratio for use for more than 6 months = 7.5; 95% CI = 1.7, 32). The association was stronger with longer duration of use compared with shorter duration of use and was more pronounced in recent users than in remote users. An unexpectedly high (11%) number of patients with secondary pulmonary hypertension had used anorexigens. The high prevalence of anorexigen use in patients with secondary pulmonary hypertension also raised the possibility that these drugs precipitate pulmonary hypertension in patients with underlying conditions associated with secondary pulmonary hypertension.

The age-adjusted mortality rates from primary pulmonary hypertension in the years immediately preceding the use of fen–phen were not different from those reported during the years of widespread use among patients aged 20–54 years. This analysis failed to support the hypothesis that the widespread use of fen–phen in the years 1992–1997 increased the incidence of primary pulmonary hypertension. “If the use [of fen–phen] during these years created an epidemic of primary pulmonary hypertension, as some have declared, such an epidemic is not reflected in the mortality database maintained by CDC” [ ].

Prognosis . Of 62 patients (61 women) exposed to fenfluramine compared with 125 sex-matched patients with primary pulmonary hypertension, 33 had used dexfenfluramine alone, 7 had used fenfluramine alone, and 5 had used both [ ]. In 17 cases fenfluramines were taken with amphetamines. Most of the patients (81%) had taken fenfluramines for at least 3 months. The interval between the start of therapy and the onset of dyspnea was 49 months (range 27 days to 23 years). The two groups differed significantly in terms of age and body mass index. Both groups had similar severe baseline hemodynamics, but the percentage of responders to an acute vasodilator was higher in patients with primary pulmonary hypertension. Hence, more patients with primary pulmonary hypertension were treated with oral vasodilators, and long-term epoprostenol infusion was more often used in fenfluramine users. Overall survival was similar in the two groups, with a 3-year survival rate of 50%.

Mechanism and pathophysiology . The mechanism of fenfluramine-associated pulmonary hypertension has been reviewed [ ]. Since only a minority of patients exposed to fenfluramines develop pulmonary hypertension, it has been postulated that a subset may be genetically susceptible. Whether there is a related genetic abnormality in the familial PPH gene located on chromosome 2q [ ] or an abnormality of the angiotensin-converting enzyme gene [ ] has yet to be explored.

Anorexigenic drugs accumulate in the lung and other tissues, especially in cellular organelles with an internal acid pH, such as lysosomes, where they bind to acidic enzymes. Lipid enzyme inhibition, lysosomal lipidosis, and associated myeloidosis are key events in the pathological cascade [ ]. Since several stimulants and other psychotropic drugs are cationic amphiphilic compounds that accumulate in the lung, brain, and other tissues, the variety of pathological mechanisms involved in the effects of this group of drugs should be kept in mind during long-term use [ , ].

The hypothesis that nitric oxide deficiency predisposes affected individuals to anorexigen-associated pulmonary hypertension has been tested in a prospective case-control comparison with two sex-matched sets of controls: patients with primary pulmonary hypertension (n = 8) and healthy volunteers (n = 12) [ ]. Lung production of nitric oxide and systemic plasma oxidation products of nitric oxide were measured at rest and during exercise, and were lower in patients with anorexigen-associated pulmonary hypertension than in patients with primary pulmonary hypertension. This deficiency may have resulted from increased oxidative inactivation of nitric oxide, as the concentrations of their oxidative products were raised in inverse proportion to nitric oxide. These findings, and earlier evidence from animal studies [ ], have given support to the hypothesis incriminating nitric oxide as a determinant of individual susceptibility to anorexigen-associated pulmonary hypertension.

The pressure response to endothelin-1 in the canine circulation has been investigated in isolated perfused dog lung [ ]. Acute treatment of the isolated lobes with fenfluramine increased pulmonary arterial pressure. Chronic treatment with fenfluramine potentiated the pulmonary vasoconstrictor response to endothelin-1. Based on these findings, the authors proposed that the pulmonary vasculature becomes hyper-reactive to vasoactive substances, such as serotonin and endothelin-1, possibly leading to pulmonary hypertension.

Anorexigen-associated severe pulmonary hypertension is clinically and histopathologically indistinguishable from idiopathic or primary pulmonary hypertension. Analysis of clonality in microdissected endothelial cells of plexiform lesions in two patients with anorexigen-associated pulmonary hypertension showed a monoclonal expansion of pulmonary endothelial cells. Accelerated growth of pulmonary endothelial cells in response to anorexigens in patients with predisposition to primary pulmonary hypertension has been speculated [ ].

Hypertension

In a few patients, hypertension was induced or aggravated by fenfluramine. The hypertension disappeared on withdrawal, but could not in all instances be reinduced by rechallenge [ ].