0
Special Report |

Anorexigens and Pulmonary Hypertension in the United States*: Results From the Surveillance of North American Pulmonary Hypertension FREE TO VIEW

Stuart Rich, MD, FCCP; Lewis Rubin, MD, FCCP; Alexander M. Walker, MD, DrPH; Sebastian Schneeweiss, MD; Lucien Abenhaim, MD
Author and Funding Information

Affiliations: *From the Section of Cardiology (Dr. Rich), Rush Medical College, Chicago, IL; the Section of Pulmonary and Critical Care Medicine (Dr. Rubin), University of California San Diego, School of Medicine, San Diego, CA; the Department of Epidemiology (Drs. Walker and Schneeweiss), Harvard School of Public Health, Cambridge, MA; and the Department of Epidemiology and Biostatistics (Dr. Abenhaim), Jewish General Hospital, McGill University, Montreal, Canada. ,  A complete list of investigators and institutions participating in the Surveillance of North American Pulmonary Hypertension is located in the Appendix.

Correspondence to: Stuart Rich, MD, FCCP, The Rush Heart Institute, Center for Pulmonary Heart Disease, Rush-Presbyterian-St. Luke’s Medical Center, 1725 West Harrison St, Suite 020, Chicago, IL 60612-3824; e-mail: srich@rush.edu



Chest. 2000;117(3):870-874. doi:10.1378/chest.117.3.870
Text Size: A A A
Published online

Background: The use of appetite suppressants in Europe has been associated with the development of primary pulmonary hypertension (PPH). Recently, fenfluramine appetite suppressants became widely used in the United States but were withdrawn in September 1997 because of concerns over adverse effects.

Materials and methods: We conducted a prospective surveillance study on patients diagnosed with pulmonary hypertension at 12 large referral centers in North America. Data collected on patients seen from September 1, 1996, to December 31, 1997, included the cause of the pulmonary hypertension and its severity. Patients with no identifiable cause of pulmonary hypertension were classed as PPH. A history of drug exposure also was taken with special attention on the use of antidepressants, anorexigens, and amphetamines.

Results: Five hundred seventy-nine patients were studied, 205 with PPH and 374 with pulmonary hypertension from other causes (secondary pulmonary hypertension [SPH]). The use of anorexigens was common in both groups. However, of the medications surveyed, only the fenfluramines had a significant preferential association with PPH as compared with SPH (adjusted odds ratio for use> 6 months, 7.5; 95% confidence interval, 1.7 to 32.4). The association was stronger with longer duration of use when compared to shorter duration of use and was more pronounced in recent users than in remote users. An unexpectedly high (11.4%) number of patients with SPH had used anorexigens.

Conclusion: The magnitude of the association with PPH, the increase of association with increasing duration of use, and the specificity for fenfluramines are consistent with previous studies indicating that fenfluramines are causally related to PPH. The high prevalence of anorexigen use in patients with SPH also raises the possibility that these drugs precipitate pulmonary hypertension in patients with underlying conditions associated with SPH.

Primary pulmonary hypertension (PPH), although a rare disease, has gained attention recently due to its association with the use of anorexigens. An epidemic of PPH occurred in the 1960s in association with the use of the anorexigen aminorex fumarate in Switzerland, Austria, and Germany.12 More recently, cases of PPH have been linked to the use of fenfluramine derivatives in the United Kingdom, France, and Belgium.38 The International Primary Pulmonary Hypertension Study (IPPHS) showed a strong association between PPH and the use of appetite suppressants, primarily the fenfluramine derivatives.9

In April 1996, the US Food and Drug Administration approved dexfenfluramine for extended use as an anorexigen in the United States. At the same time, the use of fenfluramine and phentermine together (“fen-phen”) was gaining in popularity, and cases of PPH associated with these drugs soon appeared.10 Although the IPPHS had shown that the risk of developing PPH increased with the duration of use of appetite suppressants, the experience with long duration of use in Europe was limited since the indicated use of these drugs had been restricted to 3 months. Approval in the United States, however, did not carry any limitation on the duration of use of dexfenfluramine.

Given the extraordinary prevalence of obesity in the United States, and the widespread and growing use of anorexigens without a limit on their length of use, a significant increase in the number of cases of PPH seemed likely.9 Consequently, a group of pulmonary hypertension specialists initiated a voluntary collaborative survey entitled Surveillance of North American Pulmonary Hypertension (SNAP). The goal of the project was to collect clinical information on patients with pulmonary hypertension who had been referred to their institutions and to document their exposure to commonly used medications, with special emphasis on anorexigens and chemically related substances. By doing so, we would be able to compare the association between anorexigen use and the development of PPH using patients with secondary pulmonary hypertension (SPH) as a comparison group.

SNAP was a prospective surveillance study that collected information on patients diagnosed with pulmonary hypertension in 12 centers in North America (see “Appendix”). The sites were selected based on the principal investigator’s prior activity in the National Institutes of Health Registry of Primary Pulmonary Hypertension,11 as well as their participation in recent clinical trials for the therapy of pulmonary hypertension.

Every patient who presented to these medical centers with a diagnosis of pulmonary hypertension that was confirmed was eligible for the study. The pulmonary hypertension was characterized by the investigator as primary or associated with another condition. Patient enrollment began on September 1, 1996, and ended on December 31, 1997. Each center had local institutional review board approval.

Pulmonary hypertension was defined as a mean pulmonary artery pressure, measured at cardiac catheterization, that was > 25 mm Hg or as a pulmonary artery systolic pressure, estimated by echocardiography or measured at catheterization, that was > 35 mm Hg. Patients were classified as having SPH if they had precapillary pulmonary hypertension and the coexistence of COPD, sleep apnea, pulmonary embolism, interstitial lung disease, liver disease, congenital heart disease or connective tissue disease, all of which were operationally defined for the study. Patients with pulmonary hypertension associated with the HIV were categorized separately. Patients with pulmonary hypertension that was not secondary and was not associated with HIV were categorized as PPH. Patients with evidence of left heart myocardial or valvular heart disease were excluded.

Patients were interviewed with respect to risk factors and drug exposure that were possibly related to their pulmonary hypertension, with special attention paid to the use of antidepressants, anorexigens, and amphetamines. Using standardized forms, study coordinators who had participated in a common training session recorded the patient’s drug exposure in the 10 years preceding presentation, noting the date of onset, the date last used, and the cumulative months of use. For the most part, interviews took place before the final diagnosis had been established.

Drug use was classified according to chemical class. The fenfluramine appetite suppressants (dexfenfluramine and fenfluramine racemate) were gathered together for analytic purposes, as were amphetamines, and antidepressants. Phentermine was kept as a distinct entity in the analysis. A class of “other anorexigens” included phendimetrazine, mazindol, diethylpropion, and oral phenylpropanolamine preparations. Nonprescription medications for conditions other than weight loss were not categorized or analyzed. The duration of use and the interval prior to the onset of symptoms were examined only in very broad categories (< 6 months vs ≥ 6 months) to account for anticipated errors in patient recall.

Associations between drug use and PPH vs SPH were explored through logistic regression, with adjustment for age, gender, body mass index (BMI), date of the onset of symptoms, and study center. Cases with missing data for any covariate were omitted from the regression analysis.

There were 579 patients with pulmonary hypertension who met entry criteria. Two hundred five had PPH, 367 had SPH, and 7 had pulmonary hypertension associated with HIV infection. Patients with HIV were set aside. Table 1 shows characteristics of the remaining patients, separated by diagnosis. The PPH patients were younger than the SPH patients (median age, 47 vs 53 years, respectively) and were more likely to be women (female to male ratio, 4.1:2.6). The PPH and SPH groups essentially were identical in BMI, New York Heart Association (NYHA) functional class at presentation, and pulmonary artery pressures as measured by echocardiograph and by cardiac catheterization.

Table 2 displays reported drug use by categories. Relative to patients with SPH, patients with PPH reported small, nonsignificant excesses in the use of amphetamines, anorexigens overall, and phentermine in particular. Differences in the use of fenfluramine and dexfenfluramine were more striking (PPH patients, 11.2%; SPH patients, 4.9%), particularly in patients with ≥ 6 months of usage (PPH patients, 6.8%; SPH patients, 1.1%). The recent use of fenfluramines also was more frequent in PPH vs SPH cases (7.8% vs 1.6%).

With the exception of the use of phentermine, adjusted odds ratios (ORs) from the logistic regression (Table 3 ) reflected the directions of associations present in the tabular analysis. Only the use of fenfluramines for ≥ 6 months remained strongly associated with the diagnosis of PPH as opposed to SPH (OR, 7.5; 95% confidence interval [CI], 1.7 to 32.4). When only recent users of fenfluramines (ie, those using them in the 6 months preceding diagnosis) were counted as exposed, the associated ORs were higher. For short duration of use (less than 6 months), the OR was 2.9 (95% CI, 0.7 to 12.6), while the OR for longer duration of use was 11.5 (95% CI, 1.9 to 67.7) (data not shown in Table 3).

The SNAP results show a clear association between the use of fenfluramines and the diagnosis of PPH as compared to SPH, with the association being stronger with longer duration of use. This is the first study in the United States to document an association between fenfluramine use and PPH, and it supports the IPPHS result.9

The SNAP registry was assembled on a voluntary basis with no funding support. Although the participating investigators were asked repeatedly to report all cases of pulmonary hypertension, whether they were primary or secondary cases or whether patients had been exposed to anorexigens or not, no audit was performed. It is conceivable that patients exposed to anorexigens were more likely than those not exposed to have PPH diagnosed or reported to SNAP once they had received the diagnosis. But this bias would have affected those using phentermine as well as fenfluramines, since little distinction had been made between these compounds in either the medical or lay press. Preferential diagnosis or referral in anorexigen users, in general, cannot explain an association that is unique to fenfluramines.

The use of SPH patients as comparators in the analysis compensated for diagnostic and referral bias, because the presenting symptoms of SPH patients are identical to those of PPH patients. Although referral for the evaluation of pulmonary hypertension might well have been affected by physician knowledge of anorexigen use, it is difficult to conceive of a mechanism whereby there might have been a differential referral rate according to final diagnostic status. Moreover, the patients in SNAP were for the most part very ill. More than 72% of PPH patients had severe dyspnea (NYHA classes III or IV). This is inconsistent with the possibility that known anorexigen use precipitated early referral or diagnosis. Recall bias could have occurred if PPH patients were more likely to report exposure to anorexigens than SPH patients, but this would have affected all anorexigens to the same extent. Misclassification of exposure could have occurred due to a lack of accuracy in the definition of the onset of the disease process, but we used only very broad categories of time, and this would not in any case have introduced a specific bias related to long-term fenfluramine exposure. Potential confounding by age, gender, BMI, center, and calendar time (onset of symptoms) was controlled for in the adjusted analysis.

There were reported to be approximately 6 million people exposed to diet pills during the period of the SNAP study. Although the unexpectedly high reported rate of anorexigen use in SPH patients (11.4%) could be an artifact, it could also point to a wider concern. There may have been a general, preferential referral of exposed patients with pulmonary hypertension to tertiary centers, and, therefore, to SNAP during the study period, but the advanced clinical state of the SNAP patients makes early referral highly unlikely. Nonetheless, even the possibility of preferential referral underscores the importance of using SPH as a comparator for PPH. Control subjects sampled from the general population, by contrast, would be an inappropriate comparison group.

We questioned initially whether anorexigens increase the likelihood of developing pulmonary hypertension in patients with another condition known to cause SPH. This is certainly a plausible explanation of the high prevalence of anorexigen use in the SPH group. The possible contributory effect of anorexigens to the development of SPH would be a new observation and would suggest that multiple risk factors for SPH have additive effects.

The study was conducted between September 1996 and December 1997, a time when diet pill use was markedly increasing in the United States. The withdrawal of dexfenfluramine and fenfluramine from the worldwide market in September 1997 also caused a major drop in the use of phentermine, and, fortunately, the exposure of large populations to long durations of use of the appetite suppressants did not have time to take place.

The fact that the OR for the short duration of use of fenfluramines and PPH vs SPH lacked statistical significance should not be interpreted as meaning that a short duration of exposure does not confer increased risk. PPH from fen-phen use has been reported to occur with as little as 23 days of exposure.10 The OR of 1.3 associated with the short-term use of fenfluramines by patients in SNAP points to small increases in risk at limited exposure and is consistent with the observation made in IPPHS of increasing risk with increasing duration of use.

The study does not reflect the total number of cases of PPH attributable to appetite suppressants in the United States. The 12 centers that participated in the study include the major referring centers for PPH in the country, but a large proportion of cases are diagnosed outside of these centers. In addition, since there is often a long delay between the onset of symptoms and diagnosis, many cases of PPH from diet pills have been diagnosed since this study ended.

In summary, this study supports the findings of previous reports that fenfluramines are risk factors for PPH. Their quick withdrawal from the United States market may well have aborted an incipient epidemic in the United States, all the more since the risk of developing PPH increases with the duration of use of fenfluramines. These anorexigens also may have precipitated pulmonary hypertension in patients with underlying conditions associated with SPH.

SNAP Participants
Principal Investigators

Lucien Abenhaim, MD, McGill University, Montreal, Canada; Stuart Rich, MD, Rush Medical College, Chicago, IL; and Lewis Rubin, MD, University of California, San Diego, CA.

Scientific Consultants

Yola Moride, PhD, Department of Epidemiological Biostatistics, McGill University, Montreal, Canada; Alexander M. Walker, MD, DrPH, Department of Epidemiology, Harvard School of Public Health, Cambridge, MA; and Sebastian Schneeweiss, MD, Department of Epidemiology, Harvard School of Public Health, Cambridge, MA.

Data and Coordinating Center

The Rush Heart Institute Center for Pulmonary Heart Disease, Chicago, IL.

Participating Investigators and Institutions

Robyn Barst, MD, Columbia Presbyterian Medical Center, New York, NY; Victor Tapson, MD, Duke University, Durham, NC; Bruce Brundage, MD, Harbor UCLA Medical Center, Los Angeles, CA; David Langleben, MD, Jewish General Hospital, Montreal, Quebec; C. Greg Elliott, MD, LDS Medical Center, Salt Lake City, UT; Michael McGoon, MD, Mayo Clinic, Rochester, MN; Stuart Rich, MD, Rush Heart Institute, Chicago, IL; Richard Channick, MD, University of California, San Diego, CA; David Badesch, MD, University of Colorado, Denver, CO; Lewis Rubin, MD, University of Maryland, Baltimore, MD; Harold Palevsky, MD, University of Pennsylvania, Philadelphia, PA; and James Loyd, MD, Vanderbilt University, Nashville, TN.

Abbreviations: BMI = body mass index; CI = confidence interval; IPPHS = International Primary Pulmonary Hypertension Study; OR = odds ratio; PPH = primary pulmonary hypertension; SNAP = Surveillance of North American Pulmonary Hypertension; SPH = secondary pulmonary hypertension

Table Graphic Jump Location
Table 1. Patient Characteristics in Relation to Diagnosis*
* 

Values given as median (interquartile range) or No. of patients (%). PA = pulmonary artery.

 

For approximately 6% of patients in each group, NYHA class was not recorded.

Table Graphic Jump Location
Table 2. Use of Medications Prior to First Symptom*
* 

Values given as No. of patients (%).

 

Five or fewer patients in each diagnostic group recorded the following anorexigens: phendimetrazine, mazindol, diethylpropion, and phenylpropanolamine.

 

Duration of use was not determined for one SPH patient. Patients with uncertain dates for onset of symptoms are not included.

Table Graphic Jump Location
Table 3. Adjusted ORs for the Association Between PPH vs SPH and Patient Characteristics
* 

From a logistic regression that also included terms for age, date of onset of symptoms, and center.

 

OR per 10-unit increase.

Gurtner, HP (1985) Aminorex and pulmonary hypertension.Cor Vasa27,160-171. [PubMed]
 
Kay, JM, Smith, P, Heath, D Aminorex and the pulmonary circulation.Thorax1971;26,262-269. [PubMed] [CrossRef]
 
Douglas, JG, Munro, JF, Kitchin, AH, et al Pulmonary hypertension and fenfluramine.BMJ1981;283,881-883. [PubMed]
 
Pouwels, HM, Smeets, JL, Cheriex, EC, et al Pulmonary hypertension and fenfluramine.Eur Respir J1990;3,606-607. [PubMed]
 
McMurray, J, Bloomfield, P, Miller, HC Irreversible pulmonary hypertension after treatment with fenfluramine [letter].BMJ1986;293,51-52
 
Watters, K, Le Ridant, A Irreversible pulmonary hypertension after treatment with fenfluramine [letter]. BMJ. 1986;;292 ,.:1137
 
Brenot, F, Herve, P, Petitpretz, P, et al Primary pulmonary hypertension and fenfluramine use.Br Heart J1993;70,537-541. [PubMed]
 
Delcroix, M, Kurz, X, Walckiers, D, et al High incidence of primary pulmonary hypertension associated with appetite suppressants in Belgium.Eur Respir J1998;12,271-6. [PubMed]
 
Abenhaim, L, Moride, Y, Brenot, F, et al Appetite-suppressant drugs and the risk of primary pulmonary hypertension.N Engl J Med1996;335,609-616. [PubMed]
 
Mark, EJ, Patalas, ED, Chang, HT, et al Fatal pulmonary hypertension associated with short-term use of fenfluramine and phentermine.N Engl J Med1997;337,602-606. [PubMed]
 
Rich, S, Dantzker, DR, Ayres, SM, et al Primary pulmonary hypertension: a national prospective study.Ann Intern Med1987;107,216-223. [PubMed]
 

Figures

Tables

Table Graphic Jump Location
Table 1. Patient Characteristics in Relation to Diagnosis*
* 

Values given as median (interquartile range) or No. of patients (%). PA = pulmonary artery.

 

For approximately 6% of patients in each group, NYHA class was not recorded.

Table Graphic Jump Location
Table 2. Use of Medications Prior to First Symptom*
* 

Values given as No. of patients (%).

 

Five or fewer patients in each diagnostic group recorded the following anorexigens: phendimetrazine, mazindol, diethylpropion, and phenylpropanolamine.

 

Duration of use was not determined for one SPH patient. Patients with uncertain dates for onset of symptoms are not included.

Table Graphic Jump Location
Table 3. Adjusted ORs for the Association Between PPH vs SPH and Patient Characteristics
* 

From a logistic regression that also included terms for age, date of onset of symptoms, and center.

 

OR per 10-unit increase.

References

Gurtner, HP (1985) Aminorex and pulmonary hypertension.Cor Vasa27,160-171. [PubMed]
 
Kay, JM, Smith, P, Heath, D Aminorex and the pulmonary circulation.Thorax1971;26,262-269. [PubMed] [CrossRef]
 
Douglas, JG, Munro, JF, Kitchin, AH, et al Pulmonary hypertension and fenfluramine.BMJ1981;283,881-883. [PubMed]
 
Pouwels, HM, Smeets, JL, Cheriex, EC, et al Pulmonary hypertension and fenfluramine.Eur Respir J1990;3,606-607. [PubMed]
 
McMurray, J, Bloomfield, P, Miller, HC Irreversible pulmonary hypertension after treatment with fenfluramine [letter].BMJ1986;293,51-52
 
Watters, K, Le Ridant, A Irreversible pulmonary hypertension after treatment with fenfluramine [letter]. BMJ. 1986;;292 ,.:1137
 
Brenot, F, Herve, P, Petitpretz, P, et al Primary pulmonary hypertension and fenfluramine use.Br Heart J1993;70,537-541. [PubMed]
 
Delcroix, M, Kurz, X, Walckiers, D, et al High incidence of primary pulmonary hypertension associated with appetite suppressants in Belgium.Eur Respir J1998;12,271-6. [PubMed]
 
Abenhaim, L, Moride, Y, Brenot, F, et al Appetite-suppressant drugs and the risk of primary pulmonary hypertension.N Engl J Med1996;335,609-616. [PubMed]
 
Mark, EJ, Patalas, ED, Chang, HT, et al Fatal pulmonary hypertension associated with short-term use of fenfluramine and phentermine.N Engl J Med1997;337,602-606. [PubMed]
 
Rich, S, Dantzker, DR, Ayres, SM, et al Primary pulmonary hypertension: a national prospective study.Ann Intern Med1987;107,216-223. [PubMed]
 
NOTE:
Citing articles are presented as examples only. In non-demo SCM6 implementation, integration with CrossRef’s "Cited By" API will populate this tab (http://www.crossref.org/citedby.html).

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging & repositioning the boxes below.

Find Similar Articles
CHEST Journal Articles
PubMed Articles
  • CHEST Journal
    Print ISSN: 0012-3692
    Online ISSN: 1931-3543