Article Text
PDF
Abstract
Aims We hypothesised that deprivation might represent a barrier to attain an ideal cardiovascular health (CVH) as defined by the American Heart Association (AHA).
Methods and results The baseline data of 8916 participants of the Paris Prospective Study 3, an observational cohort on novel markers for future cardiovascular disease, were used. The AHA 7-item tool includes four health behaviours (smoking, body weight, physical activity and optimal diet) and three biological measures (blood cholesterol, blood glucose and blood pressure). A validated 11-item score of individual material and psychosocial deprivation, the Evaluation de la Précarité et des Inégalités dans les Centres d'Examens de Santé—Evaluation of Deprivation and Inequalities in Health Examination centres (EPICES) score was used. The mean age was 59.5 years (standard deviation 6.2), 61.2% were men and 9.98% had an ideal CVH. In sex-specific multivariable polytomous logistic regression, the odds ratio (OR) for ideal behavioural CVH progressively decreased with quartile of increasing deprivation, from 0.54 (95% CI 0.41 to 0.72) to 0.49 (0.37 to 0.65) in women and from 0.61 (0.50 to 0.76) to 0.57 (0.46 to 0.71) in men. Associations with ideal biological CVH were confined to the most deprived women (OR=0.60; 95% CI 0.37 to 0.99), whereas in men, greater deprivation was related to higher OR of intermediate biological CVH (OR=1.28; 95% CI 1.05 to 1.57 for the third quartile vs the first quartile).
Conclusions Higher material and psychosocial deprivation may represent a barrier to reach an ideal CVH.
Trial registration number NCT00741728.
Statistics from Altmetric.com
Given increasing prevalence of obesity, diabetes and smoking (in women),1 ,2 preventing the development of traditional risk factors has been emphasised recently by the American Heart Association (AHA) as a complimentary approach to primary prevention of cardiovascular disease (CVD).3 ,4 To this end, in 2010, the AHA has proposed a simplified 7-item tool including four health behaviours (smoking, body weight, physical activity and optimal diet) and three biological measures (blood cholesterol, blood glucose and blood pressure) to promote ideal cardiovascular health (CVH).4 The goal of this primordial prevention approach is by 2020 to increase the prevalence of ideal CVH in the American population by 20% while reducing death from CVD and stroke by 20%.4 However, the prevalence of ideal CVH remains far behind the initial goals in both American and European populations.5–9 It is therefore urgent to identify barriers to the attainment of ideal CVH. Individual or contextual factors may shape an individual's capability in this regard. We have recently demonstrated that depressed individuals had substantially reduced likelihood of having ideal CVH.10 Here we hypothesise that deprivation, which might concern 14% of the French population (http://www.insee.fr), may also curtail the individuals' capability to attain ideal CVH.
A socioeconomic gradient in the distribution of cardiovascular risk factors and disease is well established.11–14 However, most studies have investigated cardiovascular risk factors separately, although they usually cluster. They also used a single marker of socioeconomic deprivation, although multifactorial indexes such as the Towsend or the Carstairs indexes15–17 exist. However, these latter define deprivation mostly from a socioeconomic perspective and measure it at an ecological (ie, by zone of residence) and not at the individual level. Beyond its socioeconomic characteristics, deprivation also encompasses psychosocial dimensions such as social networks and support,18 although their influence on ideal CVH has not yet been evaluated.
To the best of our knowledge, no community-based study has investigated the association between multidimensional aspects of deprivation at the individual level and ideal CVH. Accordingly, we aimed to quantify the (inverse) association between a validated, multidimensional index of material and psychosocial deprivation at the individual level,18 the Evaluation de la Précarité et des Inégalités dans les Centres d'Examens de Santé—Evaluation of Deprivation and Inequalities in Health Examination centres (EPICES) score and ideal CVH, as defined by the AHA criteria.4 In particular, we a priori decided to explore the possibility of a graded relationship between increasing deprivation and ideal CVH. Furthermore, because women are more often in ideal CVH than men,5–9 we additionally explored the possibility of some gender heterogeneity in the relationship between deprivation and ideal CVH.
To this end, we performed a cross-sectional analysis using the baseline data of almost 9000 men and women participating in the Paris Prospective Study 3 (PPS3).19
Methods
Study population
The PPS3 is an ongoing prospective observational cohort on novel markers for future CVD.19 It was registered in the World Health Organisation (WHO) International Clinical Trial Registry Platform (NCT00741728) in 25 August 2008. Our study complies with the Declaration of Helsinki and all the volunteers were recruited after signing an informed consent form. Between June 2008 and June 2012, 10 157 men and women aged 50–75 years were recruited at a large preventive medical centre, the Centre d'Investigations Préventives et Cliniques (IPC), in Paris (France). The IPC is a preventive medical centre that is subsidised by the French National Insurance System for Salaried Workers (Sécurité Sociale-CNAMTS). It offers a free medical examination every 5 years to all working and retired employees and their families. It is one of the largest medical centres of this kind in France, having carried out approximately 20 000–25 000 health examinations per year since 1970 for people living in the Paris area covering 11 millions of inhabitants (Paris and suburbs). The standard health check-up includes a complete clinical examination, including measurement of height, weight and blood pressure, coupled with standard biological tests after an overnight fast. A self-administered questionnaire provides information related to professional activity, lifestyle (tobacco and alcohol consumption, physical activity, diet), personal and family medical history, current health status and use of medications.20
Medication
Participants were asked to come to the IPC with either their most recent medical prescriptions and/or with their medication packages. Current medication consumption was checked by a medical doctor from the IPC in a face-to-face interview; these data were coded using the WHO Anatomical Therapeutic Chemical (ATC) classification.
Ideal cardiovascular health
Ideal CVH metrics
Each of the seven metrics of CVH was categorised as poor, intermediate and ideal using the AHA criteria.4 Only the ideal levels are defined below and illustrated in online supplementary figure S1, whereas the complete definition of each metric is detailed in the online supplementary material.
supplementary material
supplementary figures
Ideal body mass index
It corresponded to values between 18 and 25 kg/m2.
Ideal smoking status
It corresponded to never smokers or ex-smokers who stopped over 12 months or more.
Ideal blood pressure
Ideal blood pressure corresponded to untreated values <120/80 mm Hg, respectively.
Ideal blood total cholesterol and glucose
Standard biological tests were performed after an overnight fast. Ideal blood total cholesterol and ideal blood glucose corresponded to untreated values <5.18 mmol/L and untreated values <5.55 mmol/L, respectively.
Ideal physical activity
In the general self-administered questionnaire, participants who reported walking at least 1 hour every day or who were practising sports 3 times or more per week were categorised as having ideal physical activity.
Ideal diet
Habitual food intake was evaluated using an adapted version of a self-reported food frequency questionnaire, the Nouvel Auto Questionnaire Alimentaire - New self-reported questionnaire on habitual food intake (NAQA) (18 item), designed for use in epidemiological studies and that has been validated against a dietetic interview which used the historic diet method.21 Data on dietary fibre intake were not available, so that the healthy diet metric was based on intake of fruits–vegetables, fish, sugar and sodium. Participants consuming (1) vegetables and fresh fruits more than 4 times a week, (2) fish twice or more a week, (3) salt less than 1500 mg per day and (4) who never drink any sugar-sweetened beverages were categorised as ideal for this metric.
Global, behavioural and biological CVH
The definitions of global, behavioural and biological CVH status are summarised in online supplementary figure S2. According to the AHA criteria and previous studies, participants who had 0–2, 3–4 and 5–7 metrics at the ideal level were defined as having poor, intermediate and ideal global CVH, respectively.4 ,10 Furthermore, those with 0–1, 2 and 3–4 health behaviour metrics at the ideal level were defined as having poor, intermediate and ideal behavioural CVH. Those with 0–1, 2 and 3 biological metrics at the ideal level were defined as having poor, intermediate and ideal biological CVH. Moreover, participants with 5 or more metrics at the ideal level but who were otherwise under medications or had prevalent CVD were categorised as having an intermediate global CVH; otherwise they were categorised as having poor global CVH.4
The material and psychosocial deprivation index: the EPICES score
The 11 items of the EPICES deprivation score are reported in the online supplementary material. The EPICES score estimates material and psychosocial deprivation of individuals and has been used routinely since 2002 in these preventive health centres. It has been validated in France in a sample of nearly 200 000 persons against two well-known indices of deprivation, including the Townsend and the Carstairs indexes.18 An initial 42-item score covering several markers of deprivation such as nationality, employment status, income, education, occupational position, householder status, family structure, social networks, financial difficulties, stressful life events, perceived health and access to care was reduced to an 11-item score that explains 90.7% of the variance after factorial analysis. The EPICES score includes marital status (one item), health insurance coverage (one item), socioeconomic status (three items), family support (three items) and leisure and recreational activities (three items). A positive response to an item was attributed a weight corresponding to the regression coefficient, whereas a negative response was attributed a weight of 0. The score was obtained by adding each weight to the intercept and varied from 0 to 100: the higher the score, the more deprived was the subject.18 ,22 ,23 Sex-specific quartiles of the EPICES score were computed to test the hypothesis of a gradient between higher deprivation and lower odds of ideal CVH.
Perceived health
Perceived health, a powerful indicator of global health, was self-rated by the participants on a Likert scale ranging from 0 (poorest perceived health) to 10 (best perceived health).
Depressive symptoms status
As previously reported, a high level of depressive symptoms status was defined by a score ≥7 on the 13-item Questionnaire of Depression 2nd version, Abridged (QD2A) or the use of antidepressants.10 ,24
Statistical analysis
All statistical analyses were undertaken separately in men and women. The baseline characteristics of study participants were compared by quartiles of the EPICES score using Pearson χ2 test or one-way analysis of variance (ANOVA) where appropriate. The odds ratios (ORs) of each upper quartile of the EPICES score (main exposure variable) for intermediate and ideal CVH (outcome) were estimated by polytomous logistic regression, using the first quartile of the EPICES score and poor CVH as the respective reference groups. Participants with missing data on individual metrics were excluded from analysis for that metric. However, for the analysis of global CVH and of its behavioural and biological components, they were not systematically excluded when the available information on the other metrics was sufficient to assign a global, behavioural or biological CVH status. For example, a participant with 5 metrics at the ideal level and missing data on 2 metrics had the possibility of remaining in the analysis as they could be categorised as having an ideal global CVH. In contrast, in those with data on only 2 metrics, a global CVH status could not be assigned and the participant was excluded from analysis.10 Regression models were adjusted for a priori defined confounding factors such as age, alcohol consumption, depressive symptoms status, education and self-perceived health10 not present in the EPICES score. All statistical analyses were two-tailed and used a p value of <0.05 to signify statistically significant associations. Statistical analyses were performed using SAS V.9.4 (Statistical Analysis System, Cary, North Carolina, USA).
Results
Study population
Among the 10 157 subjects enrolled in the PPS3, 756 could not be assigned an EPICES score, 116 had missing data on covariates and 369 on CVH status, leading to a study population of 8916 men and women. Subjects with missing data were slightly older, more often women and had worse psychosocial and clinical characteristics than those included in the analyses (see online supplementary table S1).
Participants' general characteristics
The mean age (standard deviation, SD) was 59.48 years (SD 6.24), 61.16% were men and 2.05% had prevalent CVD. Overall, 9.98% had ideal CVH. Women compared with men were slightly older, less educated, more often depressed, in lower occupational status and reported lower perceived health; however, they were twice more often in ideal CVH (14.98% vs 6.87%, p<0.0001) (see online supplementary table S2).
Participants' characteristics by level of deprivation
The baseline characteristics of the study participants by quartile of deprivation are shown in table 1 (for women) and table 2 (for men). The cut-off values for quartiles of deprivation were higher in women than men. The prevalence of ideal CVH was 20.52% in the least deprived and 12.15% in the most deprived women; in men, this went from 7.80% to 5.67%. The burden of cardiovascular risk factors increased in a graded manner with higher level of deprivation in both genders, except for total cholesterol level and lipid-lowering drugs (in both genders), diabetes (in women) and systolic blood pressure (in men).
Characteristics of Paris Prospective Study 3 study participants by level of deprivation in women
Characteristics of Paris Prospective Study 3 study participants by level of deprivation in men
Association of deprivation with each metric of CVH
Table 3 presents the multivariable association of each increasing quartile of deprivation (compared with the first) with intermediate and ideal level for individual metric of CVH (compared with poor CVH). In both genders, higher deprivation was associated with significant and substantially lower odds for nearly all ideal behavioural metrics (except the diet metric in men). The magnitudes of these associations were similar in men and women. For the biological metrics, higher deprivation tended unexpectedly to be related to higher odds of ideal total cholesterol in both genders. In women, higher deprivation was also associated with substantially lower odds for ideal blood pressure.
Odds ratios (ORs) of higher deprivation for intermediate and ideal level of each metric of cardiovascular health (CVH) in women and men
Association of deprivation with behavioural, biological and overall CVH
The multivariable associations of higher deprivation with ideal behavioural, biological and global CVH are shown in figure 1A (in women) and figure 1B (in men), with further details in online supplementary table S3. The OR for ideal behavioural CVH progressively decreased with each quartile of (increasing) deprivation from 0.54 (95% CI 0.41 to 0.72) to 0.49 (0.37 to 0.65) in women and from 0.61 (0.50 to 0.76) to 0.57 (0.46 to 0.71) in men. Associations with ideal biological CVH were confined to the most deprived women (OR=0.60; 95% CI 0.37 to 0.99), whereas in men, greater deprivation was related to higher OR of intermediate biological CVH (OR=1.24; 95% CI 1.01 to 1.52 and OR=1.28; 95% CI 1.05 to 1.57 for the second and the third vs the first quartile, respectively). Accordingly, there was a significant sex interaction between deprivation and biological CVH (p=0.001). Regarding global CVH, each quartile of (increasing) deprivation was associated with lower odds of ideal (and intermediate) CVH in a graded manner in women with ORs ranging from 0.57 (0.42 to 0.77) to 0.48 (0.35 to 0.66). In men, only the most deprived (last quartile) had significant decreased odds of ideal (and intermediate) CVH with an OR for ideal CVH of 0.62 (0.45 to 0.85). However, there was no gender interaction between the level of deprivation and the level of CVH (p=0.24).
(A) Multivariable odds ratios (ORs) of each upper quartile of deprivation for behavioural, biological and total ideal cardiovascular health (CVH) in women. *The behavioural and biological CVH was missing in 129 and 8 women, respectively. (B) Multivariable ORs of each upper quartile of deprivation for behavioural, biological and total ideal CVH in men. *The behavioural and biological CVH was missing in 135 and 10 men, respectively. ORs were obtained by polytomous regression analysis and quantify the association between the second, the third and the fourth against the first quartile of deprivation for ideal CVH. Analysis was adjusted for age, education, alcohol consumption, perceived health and depressive symptoms status. Horizontal bars represent the 95% CI limit of each OR. OR below 1 indicates that higher deprivation is associated with lower odds of ideal CVH; OR above 1 indicates that higher deprivation is associated with higher odds of ideal CVH. In each row, the ratio of the number of subjects in ideal CVH over the sample size of the second, third and fourth quartiles of deprivation score is reported.
Discussion
In this cross-sectional analysis of 8916 men and women aged 50–75 years, increasing level of deprivation was progressively and significantly associated with lower odds of ideal CVH, especially its behavioural component. Some gender heterogeneity was reported, so that the association between higher deprivation and lower ideal CVH was more consistent in women than men, that is, in a graded manner and with a larger effect size.
Three recent studies investigated the potential impact of socioeconomic aspects of deprivation on the distribution of ideal CVH.25–27 The first two studies were conducted in Bosnia Herzegovina, a transition European country and demonstrated an inverse association between education attainment and employment status with ideal CVH, whereas no association was observed with the wealth index.25 ,26 In the third study conducted in the USA using the National Health And Nutrition Examination Surveys (NHANES) data, an inverse relationship was reported between the lack of health insurance coverage and ideal CVH. However, the association was not statistically significant after adjustment for socioeconomic status in that study.27
Our study extends the results of these three previous studies on the following issues. First, we used a multidimensional measure of deprivation, reflecting current socioeconomic and psychosocial aspects rather than a single measure of deprivation. Second, it is salient that our study is set in France, which provides access to universal healthcare to all residents for decades. This minimises, although does not exclude, residual confounding by access to preventive care. Still, our results show a frank decrement in the likelihood of ideal CVH, especially the behavioural component, with increase in deprivation. It is therefore likely that in countries with less generous health systems, the inverse association between deprivation and ideal CVH might be even worse. Third, we demonstrate that the graded inverse association with ideal CVH, especially its behavioural component, already starts in the second quartile of the EPICES score. This suggests that the inability to reach ideal CVH begins at low levels of deprivation and therefore concerns a much larger segment of the population than the most deprived group « only ». This important finding carries implications regarding the size and the typology of the population to target and the means to allocate for the promotion of ideal CVH, especially its behavioural component. The international initiative launched by the WHO in 2013 aiming at reducing premature CVD death by 25% by 2025 worldwide by tackling hypertension, obesity and smoking should help strengthening primordial prevention of CVD.28 Fourth, some gender heterogeneity was reported. Association between higher deprivation and global ideal CVH was more consistent in women than men, that is, in a graded manner and with a larger effect size. Also, the association between deprivation and ideal biological CVH went in opposite directions in women and in men. Sex disparities in the susceptibility to deprivation have been already reported, although in prior studies deprivation or its proxies were measured at the ecological level.29 ,30 The true reasons behind these sex differences remain unclear however.
The higher likelihood of ideal total cholesterol levels in the most deprived men (and women) was an unexpected finding, although a lack of health insurance coverage was associated with more ideal total cholesterol in the NHANES study.27 This might reflect a spurious result or residual confounding by unmeasured metabolic traits or unmeasured diet habits.
We acknowledge the following limitations. The age range of our study population between 50 and 75 years together with the preponderance of Caucasians limit the generalisability of our study results. Being recruited in a preventive healthcare centre, it is likely that the PPS3 participants are in better health than their counterparts of a similar age in France. This may affect the prevalence of ideal CVH and deprivation but not the association between higher deprivation and CVH. Classification bias may be present in our definitions of the physical activity metric, which was based on two simple questions, and the diet one, for which data on dietary fibre intake were unavailable. As a cross-sectional analysis, we were not able to correlate changes in deprivation with changes in ideal CVH.
To conclude, in this large community-based study conducted in France, higher material and psychosocial deprivation was associated progressively with a substantially reduced likelihood of reaching ideal (and intermediate) CVH, especially its behavioural component.
Key messages
What is already known on this subject?
A socioeconomic gradient in the distribution of single cardiovascular risk factors and disease is well established. This evidence is based on single socioeconomic marker of deprivation or multivariable index measured not in individuals but at the ecological level (ie, neighbourhood deprivation).
What might this study add?
This study investigates deprivation at the individual level using a multivariable index covering socioeconomic and psychosocial dimensions of deprivation. Furthermore, cardiovascular risk factors are not studied in isolation but in combination using a 7-item tool recently developed by the American Heart Association to define ideal cardiovascular health (CVH). We demonstrate an inverse gradient between increasing level of deprivation and ideal CVH.
Gender heterogeneity was also observed as this graded inverse association was more consistent and of greater effect size in women compared with men.
How might this impact on clinical practice?
This study suggests that higher deprivation may represent a barrier to reach ideal CVH, especially its behavioural component (smoking, body weight, physical activity and optimal diet). That the inverse association between deprivation and ideal CVH already starts at low level of deprivation carries implications regarding the size and the typology of the population to target and the means to allocate for the promotion of ideal CVH.
Acknowledgments
We thank H Khettab, N Estrugo, S Yanes, JF Pruny and J Lacet Machado for performing the study recruitment of PPS3 study participants, Dr MF Eprinchard, Dr JM Kirzin and all the medical and technical staff of the IPC Centre, the Centre de Ressources Biologiques de l'Hôpital Européen Georges Pompidou staff (C. de Toma, B. Vedie) and the Platform for Biological Resources (PRB) of the Hôpital Européen Georges Pompidou for the management of the biobank. The PPS3 is organised under an agreement between INSERM and the IPC Centre and between INSERM and the Biological Research Centre at the Européen Georges Pompidou Hospital, Paris, France. We thank the Caisse Nationale d'Assurance Maladie des Travailleurs Salariés (CNAM-TS, France) and the Caisse Primaire d'Assurance Maladie de Paris (CPAM-P, France) for helping make this study possible.
References
Footnotes
Contributors Design of the study: JPE and XJ. Drafting the manuscript: JPE. Statistical analysis: JPE and MCP. Critical review of important intellectual content: all authors. Responsibility of the integrity of the data: JPE and XJ.
Funding The PPS3 was supported by grants from The National Research Agency (ANR), the Research Foundation for Hypertension (FRHTA), the Research Institute in Public Health (IRESP) and the Region Ile de France (Domaine d'Intérêt Majeur). EWwas supported by a grant from GESTES/Région Île-de-France. BGwas supported by a grant from the French Ministry of Research and Education.
Competing interests CL has received advisory panels or lecture fees from AstraZeneca, Bristol-Myers Squibb, Lundbeck, Pfizer, Pierre Fabre, Sanofi and Servier. AS-M receives research support from the National Institute of Health (NIH) (NIA R01AG013196 (principal investigator), National Institute of Aging (NIA) R01-AG034454 (principal investigator)) and the British Medical Research Council (MRC) (G0902037 (co-investigator)) outside the submitted work.
Patient consent Obtained.
Ethics approval WHO International Clinical Trial Registry Platform: NCT00741728. The study protocol was approved by the ethics committee of the Cochin Hospital (Paris, France).
Provenance and peer review Not commissioned; externally peer reviewed.