Patients

Marketing authorization for benfluorex was suspended in France in November 2009 and the French drug regulatory agency (Agence française de sécurité sanitaire des produits de santé [Afssaps]) issued a public health advisory inviting all patients with previous exposure to benfluorex to contact their primary care physician. Primary care physicians further referred patients to a cardiologist for echocardiography. On 26 November 2009, the cardiology departments of all French university hospitals, leading private clinics and general hospitals were contacted by e-mail on behalf of the French Society of Cardiology and were invited to participate in this multicentre prospective echocardiography study. Investigators were asked to prospectively include all consecutive patients previously exposed to benfluorex, referred by their primary care physicians for echocardiography. Ten centres participated in this study (Centre Hospitalier Universitaire d’Amiens, Centre Hospitalier de Beauvais, Centre Hospitalier Universitaire de Bordeaux, Centre Hospitalier Universitaire de Brest, Centre Hospitalier de Compiègne, Centre Hospitalier Universitaire de Lille, Groupe Hospitalier de l’Institut Catholique de Lille, Centre Hospitalier Universitaire de Nantes, AP–HP Hôpital Saint-Antoine Paris and Centre Hospitalier Universitaire de Rennes). This observational study was approved by the non-interventional research ethics committee of the University of Picardie, Amiens, France. Oral consent was obtained from each patient.

Consecutive benfluorex-treated patients with no history of VHD referred by primary care physicians for echocardiography to the participating centres were enrolled in the present study when they had a history of 3 months or over of exposure to benfluorex. All patients exposed to benfluorex and referred for a second expert evaluation after initial echocardiography were not included in the present study. Patients exposed to drugs that could induce VHDs (rye ergot alkaloids, fenfluramine/phentermine, dexfenfluramine, pergolide) and patients with a cardiac valve prosthesis were excluded from the present study.

Demographic data, cardiovascular risk factors, presence of symptoms and echocardiography variables were recorded. Total duration of treatment by benfluorex was systematically recorded. Primary care physicians were contacted by phone at the time of the echocardiography when there was any doubt concerning exposure to SSRIs or other medications. Patients were classified in the SSRI group when they had been exposed to SSRIs for 3 months or over. Three patients in whom there was finally a doubt about exposure to SSRI were excluded from the present study.

Echocardiography

Complete echocardiography examinations on commercially available ultrasound devices were performed in each centre by experienced operators according to a standardized protocol with multiple two-dimensional views and the use of various Doppler modes. Views of the mitral, aortic, tricuspid and pulmonary valves were obtained whenever possible. To this effect, magnified video loops were recorded in parasternal long-axis views for the aortic and mitral valves, parasternal short-axis view for the pulmonary, tricuspid and aortic valves, apical views for the tricuspid, mitral and aortic valves, and subcostal view for all valves, whenever possible. All recordings were obtained with and without Doppler colour flow mapping. All echocardiography examinations were stored in digital DICOM format on digital versatile disks for subsequent off-line analysis.

All echocardiograms were independently read by two cardiologists who were experts in echocardiography and VHD and were blinded to all aspects of the patient’s history, including SSRI use, to assess heart valve morphology and regurgitation. In the event of disagreement between the two readers, a third independent expert performed a final blinded reading and provided a final grading. The severity of valve regurgitation was classified as absence or trace, mild, moderate or severe, according to the recommendations of the European Society of Echocardiography . The moderate regurgitation group was subdivided into ‘mild-to-moderate’ and ‘moderate-to-severe’ . Echocardiography features of drug-induced VHD were systematically investigated . For the mitral valve ( Fig. 1 ), these features were leaflet thickening, retraction towards the ventricular apex during systole (leaflet tenting), reduced valve mobility and/or thickening and shortening of the chordae tendineae. For the aortic valve ( Fig. 2 ), these features were systolic subtle dome-like appearance of the leaflets, valvular thickening, reduced mobility and/or incomplete diastolic coaptation resulting in a small central triangular valve hiatus during diastole in the short-axis view, with central aortic regurgitation (AR). Based on this blind qualitative analysis, echocardiography valvular features were classified as: highly suggestive of drug-induced VHD; inconclusive; and not suggestive of drug-induced VHD.

Apical long-axis two-dimensional views magnified on the mitral valve in systole (A) and diastole (B), and colour Doppler flow mapping in systole (C), showing leaflet thickening associated with thickening and shortening of the chordae tendineae and retraction of one or both leaflets with reduced valve mobility, with a typical drumstick appearance, in a patient with mitral valve abnormalities suggestive of drug-induced valvular heart disease and moderate mitral regurgitation. MR: mitral regurgitation; MV: mitral valve.

Parasternal long-axis two-dimensional views magnified on the aortic valve in systole (A) and diastole (B), and colour Doppler flow mapping in parasternal long-axis (C) and short-axis (D) views, showing central aortic regurgitation, leaflet thickening, presence of a small central triangular valve hiatus during diastole and presence of a subtle dome-like appearance of the aortic valve during systole, these abnormalities being suggestive of drug-induced valvular heart disease. AR: aortic regurgitation; AV: aortic valve.

Left ventricular end-diastolic and end-systolic diameters were obtained from parasternal long-axis views. Left ventricular ejection fraction was calculated using Simpson’s biplane method. The quality of the echocardiograms was graded semiquantitatively into four grades (1: excellent; 2: good; 3: average; and 4: poor).

Statistical analysis

Continuous variables are expressed as mean ± standard deviation or median [25th–75th percentiles] as appropriate. Categorical variables are expressed as absolute numbers and percentages. Comparisons between continuous variables were performed with Student’s t -test or the Mann-Whitney U test, as appropriate. Comparisons between categorical variables were performed using Chi-square or Fisher’s exact test, as appropriate. The association between SSRI use and left heart valve regurgitation ≥ mild or moderate and presence of heart valve abnormalities suggestive of drug-induced VHD was tested using complete instead of stepwise multiple logistic regression analysis. Age, sex, body mass index, smoking, hyperlipidaemia, hypertension, diabetes, coronary artery disease, New York Heart Association (NYHA) functional class III–IV, presence of cardiac murmur and duration of benfluorex therapy were a priori included in the multivariable model. Agreement between readers for the severity of aortic or mitral regurgitation (MR) (none or trace, mild, moderate, severe) and for aortic or mitral valve morphology (highly suggestive of drug-induced VHD; inconclusive; not suggestive of drug-induced VHD) was assessed using the kappa value. A two-tailed P value < 0.05 was deemed significant. All analyses were performed using PASW Statistics 18.0 (IBM, Inc., Bois-Colombes, France).