Population

The assoCiatiOns thérapeutiques chez L’HYpertendu suivi en cardioloGie ambulatoirE (COLHYGE) study was an observational cross-sectional epidemiological study that took place in France during 2009. Participation in this study was proposed to a representative panel of 2983 cardiologists, who were members of the professional association of the French cardiologists (CNCF). This panel represents almost 80% of all independent French cardiologists ( n = 3766). As the COLHYGE study was based on voluntary participation, only 516 cardiologists agreed to participate and 371 were truly active and included patients. The panel of 371 active cardiologists had the following characteristics: 24 (7%) were < 40 years,132 (35%) were between 40 and 50 years and 215 (58%) were > 50 years; 83 (22%) cardiologists had been in practice for < 10 years and 288 (78%) for ≥ 10 years; 81 (22%) worked in rural areas or small towns with < 20,000 inhabitants, 143 (39%) worked in medium-sized towns (20,000–100,000 inhabitants) and 147 (40%) worked in big cities with > 100,000 inhabitants; 214 (58%) had no associated professional activities and 157 (42%) were part-time employees (mostly in public hospitals or private clinics); 162 (44%) worked on their own and 209 (56%) worked in association with other physicians. Although not selected randomly, this panel of 371 cardiologists was representative of the French cardiologist population concerning these characteristics.

Each physician was asked to include all the patients identified from their practice who fulfilled the inclusion criteria, over a limited period. As the main objective was to describe the hypertensive population consulting a cardiologist in France, the inclusion criteria were broad: essential hypertensive patient; pharmacologically treated; willing to participate in the study and to give consent after provision of written and verbal information. Finally, 5798 consecutive patients were included in the study (mean number of patients per cardiologist with all data completed was 15.6 ± 7.9). The protocol was approved by institutional review committees (Conseil National de l’Ordre des Médecins, Comité Consultatif sur le Traitement de l’Information en Matière de Recherche dans le domaine de la Santé, Commission Nationale Informatique et Liberté).

For the patients, an inclusion questionnaire (completed by the physicians) contained the following data: sex; age; body weight; height; body mass index; hypertension duration; history of diabetes mellitus, dyslipidaemia, nephropathy, left ventricular hypertrophy, coronary heart disease, peripheral artery disease, stroke and atrial fibrillation; current smoking habit; and current use of antiplatelet, lipid-lowering and antidiabetic drugs. The antihypertensive treatment of the patients was detailed, in terms of number of antihypertensive drugs, use of different antihypertensive classes and use of fixed combinations, as well as previous adverse effects related to antihypertensive drugs.

BP and heart rate were measured after a 5-minute rest. Different devices were used – according to the device generally used by the practitioner (mostly electronic devices) – and the mean of three determinations was used for systolic and diastolic BP. Pulse pressure was calculated as systolic minus diastolic. Mean BP was diastolic plus a third of pulse pressure. According to the 2007 European Society of Hypertension Guidelines , based on these measurements and on questionnaires, BP was considered controlled if systolic and diastolic BPs were < 140/90 mmHg for patients without any associated high cardiovascular risk condition and < 130/80 mmHg for patients with diabetes mellitus or previous atherothrombotic cardiovascular disease or nephropathy defined by either microalbuminuria (30–300 mg/day or 20–200 mg/L), macroalbuminuria (> 300 mg/day or > 200 mg/L) or renal failure defined by a creatinine clearance < 60 mL/min.

Statistical analysis

Quantitative variables were described using means and standard deviations; qualitative variables were described by numbers and percentages. The dependent variable analysed was uncontrolled BP, according to the 2007 European Society of Hypertension Guidelines . First, each variable (patient- and physician-related) was tested in association with this dependent variable in a bivariate logistic regression analysis; second, a linear multilevel mixed logistic model was performed with all statistically significant variables in the bivariate analysis (plus age, sex and body mass index) using a backward manual selection procedure, considering P > 0.05 as the exit cut-off level. Some variables, such as mean arterial pressure and pulse pressure, were excluded from the model because of too strong an intercorrelation with uncontrolled BP. As age, sex and body mass index are well-accepted determinants of BP control, those three variables were forced in the final model. Two levels of fixed effects were considered: the patient level (level 1) and the physician level (level 2). As the presence of nephropathy was assessed in only 3029 patients, this variable was not included in the multivariable analysis. All analyses were performed using SAS software, version 8.2 (SAS Institute, Cary, NC, USA).