Summary
Background
Hypertensive patients with established cardiovascular or renal disease (ECVRD) have an added 10-year risk of cardiovascular events, classified by the European Society of Hypertension/European Society of Cardiology as ‘very high’.
Aims
To identify determinants of blood pressure (BP) outcomes in hypertensive patients with and without ECVRD treated in second-line with valsartan.
Methods
This was a subgroup analysis comparing patients with and without ECVRD who participated in the PREVIEW study, a 90-day observational prospective effectiveness study of valsartan, conducted in Belgium. Two-level (patients ‘nested’ under physicians) hierarchical linear and logistic modelling of BP values and BP control (140/90 mmHg; 130/80 mmHg for diabetics) at 90 days was applied to data from 1107 patients with and 2087 patients without ECVRD treated with valsartan by 504 general practitioners.
Results
Absolute reductions in BP were similar across subgroups, with minor variations in actual BP levels in general and by subgroup. Fewer patients with versus without ECVRD achieved targets for systolic BP, diastolic BP and combined systolic/diastolic BP control. Variability in BP values and control at 90 days attributable to a physician-level class effect ranged from 24.6% to 28.1% and 15.0% to 22.4%, respectively. Physician- and patient-related determinants of 90-day BP outcomes varied considerably between the two subgroups.
Conclusion
Several determinants of BP outcomes were identified comparing patients with and without ECVRD, including amenable physician-level and patient-level factors and warning signs for continued risk of uncontrolled BP. ECVRD patients present with differential characteristics, conditions and determinants that mandate individualized attention to complement general evidence-based antihypertensive treatment.
Résumé
Justification
Les patients hypertendus ayant une atteinte cardiovasculaire ou rénale ont une augmentation du risque d’évènement cardiovasculaire à dix ans, considérée comme élevée dans la classification de l’European Society of Hypertension/European Society of Cardiology.
Objectifs
Identifier les déterminants de la pression artérielle chez les patients hypertendus, avec ou sans attente rénale ou cardiovasculaire, traités en seconde intention par valsartan.
Méthode
Cette étude est une analyse en sous-groupes de l’étude PREVIEW, comparant des patients avec et sans atteinte rénale et cardiovasculaire participant à cette étude, avec une période d’observation de 90 jours, effectuée en Belgique. Un modèle linéaire hiérarchisé a deux niveaux : patients et médecins ainsi qu’un modèle logistique de valeurs des pressions artérielles et de son contrôle (140/90 mmHg ; 130/80 mmHg chez les diabétiques) ont été appliqués à 90 jours, parmi les 1107 patients avec et 2087 patients sans atteinte rénale ou cardiovasculaire traités par valsartan par 504 médecins généralistes.
Résultats
La réduction absolue de la pression artérielle était similaire dans les différents sous-groupes avec des variations mineures de niveaux de pression artérielle, tant dans la population générale que dans l’analyse en sous-groupes. Un nombre moindre de patients, avec versus sans atteinte rénale ou cardiovasculaire, a atteint les objectifs cibles de niveau de pression artérielle systolique, diastolique ainsi que l’association des deux. La variabilité dans les valeurs de pression artérielle à 90 jours attribuable à l’effet médecin variait de 24,6 à 28,1 % et de 15 % à 22,4 %, respectivement. Les déterminants liés aux patients ou aux médecins de l’objectif de pression artérielle à 90 jours variaient de façon significative entre les deux sous-groupes.
Conclusion
De nombreux déterminants de l’évolution de la pression artérielle ont été identifiés chez des patients avec et sans atteinte rénale et cardiovasculaire prenant en considération des facteurs liés aux médecins et aux patients, ainsi que la prise en compte de signes d’alerte pour un contrôle des niveaux de pression artérielle. Les patients ayant une atteinte rénale ou cardiovasculaire avaient des caractéristiques différentes et des déterminants également différents, justifiant une approche individualisée pour compléter le traitement antihypertenseur en fonction de la présence ou de l’absence de ce retentissement rénal ou cardiovasculaire.
Background
Cardiovascular disease resulting from poor blood pressure (BP) control is a leading cause of death globally . Up to 13.5% of premature deaths – or 7.6 million deaths worldwide annually – are attributed to suboptimal control of arterial hypertension . Europeans are among those with the highest systolic BP (SBP) in the world and thus are at higher risk of end-organ damage . Achieving guideline-recommended BP targets is difficult in general and an even greater challenge in patients with established cardiovascular and/or renal disease (ECVRD) . The European Society of Hypertension/European Society of Cardiology (ESH/ESC) guidelines for the management of hypertension classify patients with ECVRD as having a ‘very high added’ 10-year risk for (additional) cardiovascular events. Identifying determinants of BP outcomes in patients with ECVRD and how these may be different from those in patients without ECVRD is critical to effective hypertension management and prevention of further cardiovascular and renal disease.
We report here on subgroup analyses for 1107 patients with and 2087 patients without ECVRD who were included in the Belgian PREVIEW study by 504 general practitioners (GPs) . PREVIEW was a prospective pharmacoepidemiological multicentre study that examined determinants and predictors of BP reduction and control following 90-day second-line treatment with the angiotensin II receptor blocker (ARB) valsartan. The purpose of the subgroup analyses was fourfold: to examine whether there were significant differences in patient characteristics, antihypertensive treatment and BP values and control at 90 days in patients with and without ECVRD; to quantify the proportions of variance in BP outcomes at 90 days attributable to physician-level versus patient-level factors; to identify the multilevel (patients ‘nested’) determinants of BP values at 90 days in these two subsamples; and to identify independent predictors of uncontrolled BP at 90 days in both subsamples.
Methods
The methodology of the PREVIEW study is described in detail in the article reporting the results for the entire sample, as are data on the physician sample, including knowledge of and practice according to evidence-based hypertension guidelines . Elements of relevance are summarized below.
Design
PREVIEW was designed as a 90-day prospective multicentre multilevel (patients ‘nested’ under physicians) pharmacoepidemiological study of the effectiveness of valsartan-centric antihypertensive regimens and the determinants thereof, in patients in whom prior treatment had failed or was not tolerated. A total of 3194 hypertensive patients (SBP ≥ 140 mmHg and/or DBP ≥ 90 mmHg; SBP ≥ 130 mmHg and/or DBP ≥ 80 mmHg for diabetic patients) treated by 504 general practitioners in Belgium were started on a valsartan-centric regimen in accordance with their physician’s best clinical judgment and re-evaluated 90 days later.
Samples for subgroup analyses
Patients were considered to have ECVRD if they presented at baseline with at least one of the following cardiovascular and/or renal conditions comorbid to their arterial hypertension: myocardial infarction; angina pectoris; coronary revascularization; heart failure; ischaemic or haemorrhagic cerebrovascular accident; transient ischaemic attacks; intermittent claudication; peripheral bypass/stent or amputation; or diabetic nephropathy or any other nephropathies as evidenced by creatinine greater or equal to 1.5 mg/dL and/or proteinuria. The subsamples consisted of patients whose treating physician had decided independently, according to best clinical judgment and within the approved label and reimbursement regulations, to prescribe one of three formulations of valsartan (80 mg, 160 mg or 80 mg/12.5 mg hydrochlorothiazide) as second-line monotherapy or combination therapy because first-line therapy had failed or was not tolerated. Patients with known sensitivity to ARBs or hydrochlorothiazide were excluded, as were patients concomitantly treated with an ARB other than valsartan.
Data model
Physician questionnaire
The physician questionnaire comprised: practice type, location/setting and patient mix; demographics; sources of information and knowledge related to hypertension; hypertension management practices; prescription patterns; management of adverse effects; SBP/DBP thresholds for treatment initiation and intensification; perceptions of patient adherence; and knowledge of practice guidelines.
Baseline patient data
The baseline patient data acquired were as follows: demographics; hypertension and cardiovascular history; comorbidities; lifestyle; prior antihypertensive medications; SBP and DBP; clinical data; starting valsartan dose; class of concomitant antihypertensives; and patient-reported number of days they had been non-adherent in the preceding 4 weeks.
Patient follow-up data
The patient follow-up data acquired after 90 days were as follows: SBP and DBP; concomitant drug(s) taken or changed since previous visit; clinical data; patient-reported number of days they had been non-adherent in the preceding 4 weeks; changes in valsartan dose since previous visit; and adverse effects over the past 90 days.
GPs were asked to measure BP three times at 1 to 2-min intervals in a sitting position after 5 min of rest using a calibrated standard sphygmomanometer and appropriately-sized cuff placed at heart level . The mean of the three sitting measurements without rounding was recorded as the SBP and DBP.
Statistical analysis
Standard descriptive statistics were used to describe the subsamples, including proportions and appropriate measures of central tendency and dispersion. Comparisons between subsamples were made using independent sample t tests, corrected as necessary for unequal variances, or Pearson’s or Yates’s χ 2 tests (contingency corrected) or Fisher’s exact test where appropriate.
We hypothesized that BP outcomes in each subsample were related to physician- and patient-level variables. Each participating physician recruited several patients; therefore patients could not be considered independent but instead ‘nested’ under their treating physician. We assumed that the n j patients recruited by physician j might share some proportion of variance in BP values attributable to their common physician and that this physician’s influence might impact BP values prior to any patient-specific variables. Accordingly, we applied two-level hierarchical linear and logistic modelling for each subsample . The intraclass correlation coefficient (ICC) quantified the variability in patient outcome attributable to within-physician variability before any patient-level determinants were considered. Likelihood-ratio tests were performed to determine if the physician-level ICC was significantly greater than zero. Adjusted slope coefficients or odds ratios and 95% confidence intervals (CI) were calculated to estimate the direction and strength of the relationship between individual factors and BP values and control. The ESH/ESC and Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC-7) guidelines for BP control (140/90 mmHg for non-diabetics; 130/80 mmHg for diabetics) were used to determine controlled versus uncontrolled BP. Statistical significance was set at P < 0.05, with corrections for multiplicity applied as necessary.
Results
Patient characteristics, medical history and clinical status at baseline
Of the 3194 patients with evaluable data, 1107 (35.9%) had ECVRD whereas 2087 (64.1%) did not. Compared with patients without ECVRD, patients with ECVRD tended to be older, male and living alone ( Table 1 ). A larger proportion of patients with ECVRD had significant risk factors at baseline, including lack of physical exercise, dyslipidaemia, diabetes, elevated C-reactive protein and family history of early cardiovascular disease. These patients had been treated for hypertension longer, with more classes of antihypertensives and with more prior antihypertensive medication changes. Their SBP and DBP at the time of enrollment were slightly but significantly lower compared with patients without ECVRD. Both groups had statistically similar proportions of patients who admitted to having been non-adherent to their antihypertensive medication prior to enrolment in the study.
| With ECVRD ( n = 1107) | Without ECVRD ( n = 2087) | P | |
|---|---|---|---|
| Age (years) | 69.1 ± 10.2 | 60.4 ± 11.5 | < 0.001 |
| Female | 49.7 | 53.7 | 0.011 |
| Living alone | 30.6 | 22.6 | < 0.001 |
| Race | 0.256 | ||
| White | 97.7 | 98.1 | |
| Non-white | 2.3 | 2.0 | |
| Risk factors | |||
| Hypercholesterolaemia | 63.7 | 53.9 | < 0.001 |
| Diabetes mellitus | 28.8 | 15.9 | < 0.001 |
| Smoker | 25.8 | 29.9 | 0.008 |
| Excess alcohol intake | 16.8 | 17.1 | 0.435 |
| Lack of regular exercise | 62.3 | 55.1 | < 0.001 |
| Obesity | 43.5 | 45.2 | 0.197 |
| Family history of premature cardiovascular disease | 19.9 | 15.9 | 0.003 |
| C-reactive protein ≥ 1 mg/dL | 6.2 | 2.3 | < 0.001 |
| Total cholesterol mg/dL | 217.1 ± 40.6 | 219.0 ± 39.3 | 0.220 |
| HDL mg/dL | 55.2 ± 18.0 | 57.6 ±18.6 | 0.001 |
| LDL mg/dL | 129.1 ± 35.7 | 131.3 ± 35.3 | 0.124 |
| Hypertension history | |||
| Years on hypertensive therapy | 6.9 ± 6.7 | 4.8 ± 5.7 | < 0.001 |
| Number of prior antihypertensive medication switches | 1.7 ± 1.6 | 1.2 ± 1.4 | < 0.001 |
| Number of prior antihypertensive classes (of six) | 1.9 ± 0.9 | 1.6 ± 0.8 | < 0.001 |
| Not adherent with antihypertensive medications at baseline | 36.7 | 35.0 | 0.190 |
| SBP at enrollment | 153.5 ± 16.0 | 154.9 ± 15.3 | 0.023 |
| DBP at enrollment | 90.4 ± 9.2 | 91.8 ± 9.2 | < 0.001 |
| Treatment patterns | |||
| Valsartan 80 mg | 12.8 | 13.7 | 0.022 |
| Valsartan 160 mg | 68.8 | 71.8 | |
| Valsartan 80 mg + HCTZ 12.5 mg | 18.4 | 14.6 | |
| Concomitant diuretic | 35.8 | 28.1 | < 0.001 |
| Concomitant β-adrenergic blocker | 36.2 | 27.2 | < 0.001 |
| Concomitant calcium antagonist | 23.2 | 14.8 | < 0.001 |
| Concomitant α-adrenergic blocker | 3.9 | 1.9 | 0.001 |
| Concomitant ACE inhibitor | 4.8 | 2.0 | < 0.001 |
a Missing data not reported, thus totals may not equal 100%.
Evolution of BP from baseline to follow-up
At 90 days, mean SBP was not statistically different between the two groups, but mean DBP was lower among patients with ECVRD ( Table 2 ). Mean SBP and DBP at follow-up were statistically similar between diabetic patients with and without ECVRD, but were higher among non-diabetic patients without ECVRD. Mean absolute changes in mmHg from baseline to follow-up were statistically similar among patients with and without ECVRD in general and stratified by diabetic status. BP control rates at baseline were not statistically different between the ECVRD and non-ECVRD groups in general and stratified by diabetic status. At follow-up, the proportions of patients with controlled SBP, DBP and combined SBP/DBP were higher among patients without ECVRD in general, although these differences were not evident when stratified by diabetic status.
| Patients with ECVRD | Patients without ECVRD | |||
|---|---|---|---|---|
| BP values | Absolute Δ | BP values | Absolute Δ | |
| SBP (mmHg) | ||||
| All patients | 138.7 ± 12.1 | –14.8 ± 16.6 | 139.1 ± 11.9 | –15.7 ± 15.9 |
| Non-diabetics | 138.2 ± 12.0 | –15.7 ± 16.8 | 138.8 ± 11.8 a | –15.8 ± 15.7 |
| Diabetics | 139.6 ± 12.2 | –12.6 ± 15.7 | 140.9 ± 12.6 | –15.0 ± 16.8 |
| DBP (mmHg) | ||||
| All patients | 82.3 ± 7.5 | –8.2 ± 10.1 | 82.8 ± 7.3 a | –8.9 ± 9.9 |
| Non-diabetics | 82.0 ± 7.5 | –8.6 ± 10.1 | 82.9 ± 7.3 a | –9.1 ± 9.9 |
| Diabetics | 82.7 ± 7.5 | –7.4 ± 10.1 | 83.2 ± 7.4 | –7.9 ± 10.1 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
