Geographic differences in outcomes in outpatients with established atherothrombotic disease: results from the REACH Registry

Abstract

Aims

There are major differences in the prevalence and management of patients with atherothrombotic disease including coronary artery disease (CAD), cerebrovascular disease (CVD) and peripheral artery disease (PAD) across different geographical regions. There is, however, little data allowing comparisons of management and outcomes across broad geographic regions. We aimed to describe geographical differences in baseline characteristics, management and outcomes in stable outpatients with established atherothrombotic disease.

Methods and results

From the REACH Registry of atherothrombosis, patients with documented CAD, PAD or CVD and with 4-year follow-up were included. Baseline characteristics, treatments and 4-year outcomes were recorded. Event rates were compared between geographical regions and were adjusted for risk scores predicting ischemic and bleeding events. The analyses of baseline characteristics and medications according to geographical region showed marked differences. For the composite primary outcome (cardiovascular death, non-fatal myocardial infarction (MI) and non-fatal stroke), rates ranged from 12.1% in Japan to 18.2% in Eastern Europe. After adjustment, substantial variations remained: taking North America as a reference, patients from Western Europe and Japan had a lower risk of primary outcome event (hazard ratio (HR) 0.93; p = 0.045, and HR = 0.67; p < 0.001 respectively) whereas patients from Eastern Europe had a higher risk (HR = 1.24; p < 0.001). There were no obvious differences between patients from North America and those from Latin America, the Middle East and Asia.

Conclusion

There are important variations in the outcomes of patients with atherothrombotic across geographic regions. These observations have important implications for public health and clinical research.

Introduction

Atherothrombotic disease, including coronary artery disease (CAD), cerebrovascular disease (CVD) and peripheral artery disease (PAD), represents the leading cause of death around the world.¹ There are major differences in the prevalence, management and outcomes of atherothrombotic disease across different geographic regions but there is little data allowing for comparison of management and outcomes across these broad areas. Most large recent randomized controlled trials (RCTs) have enrolled patients in various regions of the world.^2,3 However, while RCTs are the cornerstone of research, due to their strict inclusion and exclusion criteria, they do not always provide a reliable picture of real world practice and outcomes and are often focused on one specific atherothrombotic disease. Conversely, registries provide an insight into the real world but they often cover a limited aspect of the disease or a limited geographical region.

The Reduction of Atherothrombosis for Continued Health (REACH) Registry enrolled outpatients covering the entire spectrum of stable atherothrombosis in a geographically diverse population encompassing five regions.^4–7

Based on data from the contemporary REACH Registry, we aimed to describe geographical differences in baseline characteristics, management and outcomes in stable outpatients with established atherothrombotic disease.

Methods

Population

The methods of the REACH Registry have been previously detailed.^4,5 Briefly, the REACH Registry recruited outpatients aged 45 years or older with either three or more risk factors for atherosclerosis, or with documented CAD, PAD or CVD over an initial 7-month recruitment period between December 2003 and June 2004. Because of regulatory requirements in Japan, the enrolment in that country occurred between August 2004 and June 2004. Only patients with documented CAD, PAD or CVD were included in the present analysis. Documented CAD consisted of one or more of the following criteria: stable angina with documented CAD, history of unstable angina with documented CAD, history of percutaneous coronary intervention (PCI), history of coronary artery bypass graft surgery, or previous myocardial infarction. Documented CVD consisted of a hospital or neurologist report with the diagnosis of transient ischemic attack or ischemic stroke. Documented PAD consisted of one or both criteria: current intermittent claudication with ankle-brachial index of less than 0.9 or a history of intermittent claudication together with a previous and related intervention, such as angioplasty, stenting, atherectomy, peripheral arterial bypass graft, or other vascular intervention including amputation.

Follow-up

Detailed baseline characteristics were collected and subsequent annual follow-up performed. Initially, 2-year follow-up was scheduled. An additional 2-year follow-up was subsequently proposed. Only sites participating in the 4-year follow-up were included in the present analysis.

Baseline characteristics were collected after inclusion. Risk factors were defined as follows: diabetes, any history of diabetes or current diabetes treated or untreated; hypertension, previous or current hypertension with antihypertensive agents; current smoking (at least five cigarettes per day as a mean within the last month before inclusion) Obesity was defined according to the WHO body mass index (BMI) cut-off points: ≥27.5 kg/m² for Asian populations and ≥30 kg/m² for non-Asian populations.⁸ As there are some discrepancies between prognostic value of BMI and waist circumference, that have been described,⁹ abdominal obesity was also analyzed and defined according to ethno-specific waist circumference thresholds:¹⁰ North America, Latin America, Western Europe, Eastern Europe ≥ 102 cm for men/ ≥ 88 cm for women; Middle East ≥ 94 cm for men ≥ 80 cm for women; Asia ≥ 90 cm for men ≥ 80 cm for women; Japan ≥ 85 cm for men ≥ 90 cm for women.

Endpoints were not adjudicated. The primary outcome for this analysis was the composite of cardiovascular death, non-fatal myocardial infarction (MI) or non-fatal stroke and was called ‘major CV events’. Cardiovascular death included death from stroke, MI or other cardiovascular death (including death from pulmonary embolism; any sudden death unless proven otherwise by autopsy; death following a vascular operation, vascular procedure or amputation; death attributed to heart failure; death following a visceral or limb infarction; and any other death that could not be definitely attributed to a nonvascular cause or haemorrhage). Any MI or stroke followed by death in the next 28 days, regardless of the cause, was considered to be a fatal MI or fatal stroke. Strokes were categorized as ischemic, haemorrhagic or unknown on the basis of available imaging. Different types of bleeding were recorded: non-fatal haemorrhagic strokes and bleeding requiring both hospitalization and transfusion.

Statistical analysis

Continuous variables are expressed as mean (standard deviation) and categorical variables are expressed as frequencies and percentages. Baseline characteristics and medications were compared between geographical regions using analysis of variance for age and the χ² test for categorical variables. Crude event rates were determined from 4-year Kaplan-Meier estimates and were compared between geographical regions using the log-rank test. Event rates were not calculated for geographic regions with less than five events; these geographic regions were excluded from the comparison for this endpoint. For a given endpoint, deaths that were not included in the endpoint were treated as censored events. We further investigated the geographical variations in the 4-year period through the calculation of hazard ratios (HRs) using a Cox proportional hazards model adjusted for age and sex, with North America as a reference. The proportional hazards assumptions were checked using the log–log survival plots and by introducing time-dependent variables into models. Cox proportional hazards models were further adjusted for risk scores derived from the REACH Registry. For non-bleeding outcomes, Cox proportional hazards models were adjusted for the risk score predicting secondary cardiovascular events after exclusion of the geographic items of the score.¹¹ For bleeding outcomes, Cox proportional hazards models were adjusted for the bleeding risk score.¹² Adjusted events rates were calculated using the corrected group prognosis method;¹³ adjustments on age and REACH risk scores were made using the quintile values. Statistical testing was done at the two-tailed α level of 0.05. Data were analysed using the SAS software package (version 9.3; SAS Institute, Cary, NC, USA).

Results

A total of 69,055 patients were initially enrolled in 44 countries of whom 45,227 patients completed 4-year follow-up. Of these, 8073 patients had risk factors but no established disease and 37,154 patients had established atherothrombotic disease (CAD, CVD or PAD). This latter group represents the study population. Those patients were enrolled at 3647 centres in 29 countries. Although Japan belongs to Asia, it was analyzed separately because of delayed participation.

Detailed information by country and by type of disease is provided in Table e-1 (see Supplementary Material).

Baseline characteristics

Mean age was 68.2 ± 10 years and 68.0% were men. Risk factors were as follow: diabetes mellitus 36.8%, hypercholesterolemia 68.2%, obesity 27.8%, current smokers at baseline 14.8% and hypertension 79.5%. The analyses of baseline characteristics according to geographical zone showed marked differences (Table e-2). Briefly, patients enrolled in North America and in Japan were older (respectively 70.1 ± 10.3 and 70.3 ± 8.8) compared to those enrolled in the rest of Asia (65.4 ± 8), in Latin America (67.3 ± 10.1) or in Eastern Europe (60.3 ± 9.4). The proportion of patients with diabetes showed marked variations (from 25.7% in Eastern Europe to 42.2% in North America). The proportion of patients with hypercholesterolemia ranged from 43.1% in Japan to 83.7% in Middle East and the proportion of patients with obesity from 11.7% in Japan to 38.0% in North America.

The analysis of baseline medications also showed marked differences (detailed in Table e-3). The use of statins ranged from 41.2% in Japan to 87.2% in the Middle East. The use of antiplatelet agents ranged from 83.7% in North America to 94.4% in the Middle East.

Considering the subset of disease, at inclusion, 26,389 patients had CAD, 12,805 patients had CVD and 5869 had PAD (patients with several vascular beds diseased could be enrolled). Baseline characteristics according to subset of disease are shown in Table e-4, baseline medications are shown in Table e-5 and patient distribution by vascular beds and geographic regions is shown in Table e-6.

Ischemic outcomes

The analysis of unadjusted ischemic outcomes in the overall population showed marked variations (Table 1). For composite primary outcome, rates ranged from 12.1% in Japan to 18.2% in Eastern Europe. This higher risk among patients from Eastern Europe appeared to be driven by a higher risk of non-fatal stroke (8.4% compared to 4.4% in North America). Rates of 4-year CV death varied widely (3.8% in Japan to 9.5% in Latin America). Unadjusted all-cause mortality ranged from 8.5% in Japan to 15.6% in North America.

As in the overall population, when unadjusted ischemic outcomes were analysed within each atherothrombosis subset (CAD, CVD and PAD), there were marked variations (Table e-7).

After adjustment for age and sex, or for the REACH risk score,¹¹ substantial variations remained (Figure 1 and Table e-8). Taking North America as a reference, patients from Western Europe and Japan had a lower risk of recurrent event: HR = 0.93 (95% CI 0.86–1) and HR = 0.67 (95% CI 0.60–0.75) respectively whereas patients from Eastern Europe had a higher risk: HR = 1.24 (95% CI 1.14–1.36).

Similar observations were made for the subset of patients with CAD (Table e-9 and Figure 2). Among patients with CVD, the primary endpoint was not lower in Western Europe but other differences described in the overall population persisted. Among PAD patients, the only notable variation was a lower risk of the primary outcome in Japan compared to North America.

Bleeding

There were important variations in total bleeding rates (including non-fatal haemorrhagic stroke and bleeding requiring both hospitalization and transfusion) ranging from 1.8% in Eastern Europe to 3.9% in North America (Table 1).

After adjustment for age and sex, or for the REACH bleeding risk score,¹² substantial variations remained (see Figure 2 and Table e-8). There was a lower risk for bleeding in Western Europe: HR = 0.71 (95% CI 0.60–0.84); Eastern Europe: HR = 0.61 (95% CI 0.47–0.79) and Japan: HR = 0.49 (95% CI 0.38–0.64) compared to North America used as reference (Figure 2). Similar observations were made in the CAD and PAD subsets (Table e-9 and Figure 2). In the CVD population, Western Europe and Japan (but not Eastern Europe) also had lower rates of the primary outcome.

Discussion

There were major variations in patient characteristics, management and 4-year outcomes according to geographic regions, with Japan consistently emerging as a region with lower ischemic and bleeding event rates, whereas Eastern Europe appeared to have higher ischemic event rates. These differences persisted after multivariate adjustment and appeared robust across the various arterial beds considered.

While there are data regarding the prevalence of cardiovascular diseases¹⁴ or the death rate related to cardiovascular diseases¹ throughout the world, little data are available regarding geographical variations in the management and outcomes of stable atherothrombosis worldwide, particularly outside of Europe or North America. The REACH Registry provides a large robust database to explore such potential differences.

Differences in baseline characteristics and management

International differences in baseline characteristics have already been described in the global REACH cohort^4,7 and similar observations are made on the 4-year follow-up cohort: higher prevalence of obesity in North America, patients globally undertreated with antiplatelet agents and statins with large geographic variations for the latter.

International differences in outcomes

Data regarding prevalence and outcome of atherothrombotic disease are available at a national level in some high-income countries. In the United States, the CRUSADE Registry examined the in-hospital management and outcomes of Acute coronary syndrome (ACS) patients.¹⁵ International differences in outcomes are available for specific manifestations of atherothrombosis. In the field of ACS, the Global Registry of Acute Coronary Events (GRACE) Registry allowed description of in-hospital outcomes and management of ACS patients.¹⁶ The European Network for Acute Coronary Treatment (ENACT) provided a description of ACS management across Europe but no information on outcome.¹⁷ The Euro Heart Survey allowed description of baseline characteristics, management and outcomes of stable patients,¹⁸ ACS patients¹⁹ and patients treated with PCI.²⁰

To our knowledge, the REACH Registry is the first registry allowing worldwide comparison of outcomes in a population covering the full spectrum of atherothrombotic disease. Large variations between geographic regions were observed for both ischemic and bleeding outcomes. After adjustment on risk scores, significant variations remained.

Explanations for the variations

The explanations for the variations observed are probably very diverse. As reported in the description of baseline characteristics, the prevalence of risk factors varied widely between geographic regions. Differences in management are also likely to contribute to outcome variations. We observed large variations in medication use as described in Table e-3. Improvement in secondary prevention (in terms of medications as well as health education) probably has the potential to decrease the observed differences.

Country-specific factors such as economic resources, healthcare investment and health system characteristics may have also affected the observed variations in outcomes. For example, it was shown from the REACH Registry that global variations in the prevalence of elevated cholesterol among patients with history of hyperlipidemia were associated with country-level economic development and health system indices such as out-of pocket expenditures.²¹

Genetic variations in risk factor susceptibility, disease mechanisms or response to therapies may partially explain the variations observed. This aspect is still underexplored and further research is required, especially to detect genetically-based susceptibility to medications.

Beyond genetics, lifestyle factors might have a great influence. For example, the incidence of CAD in Japanese participants increased when they moved to Hawaii or California.²²

Finally, differences in disease awareness and endpoint ascertainment, such as variation in the threshold for deciding on transfusions, can also contribute to the variations observed. Although data were collected centrally via use of standardized case report form, this latter form was completed at the study visit by the treating physician.

Implications for public health and clinical research

These observations have several important implications for public health and clinical research. Rates of cardiovascular events were, as a whole, higher in low-income countries. In those countries, effort should be made to improve secondary prevention (in terms of access to medication as well as in terms of health education)²³ and financial investment could probably have an important impact.^21,24 Importantly, even in high-income regions such as Western Europe and North America, there is clearly room for improvement, as these were not the regions with the lowest event rates.

The planning of RCTs requires robust estimates of disease prevalence, and more importantly, expected incidence of major cardiovascular events by region, particularly as contemporary cardiovascular clinical trials are now, by necessity, a global endeavour.^25,26 The present observations provide a benchmark for the estimation of frequency of events for stable atherothrombotic patients in broad geographic regions which contribute to RCTs.

Limitations

These analyses are derived from an observational registry and therefore are subject to the inherent limitations of observational databases, however large these are. Even though multivariate analyses were performed with attempts to adjust for potential confounders, residual confounding cannot be ruled out. Thus, the differences observed here should be interpreted conservatively.

Some important geographic regions (such as sub-Saharian Africa) and ethnic groups did not participate in REACH. Others did participate but sometimes in numbers too low to provide robust estimates (e.g. China). In addition, even in an observational registry, there is concern that patients and physicians participating may not necessarily be representative of the overall patient and physician population and may overestimate patient compliance or physician adherence to evidence-based therapies.²⁷

Events were not adjudicated. The primary ischemic outcome includes hard endpoints (CV death, MI or stroke) that are less influenced by subjectivity than softer endpoints. However, variations in prevalence of certain events may reflect variations in management: e.g. since major bleeding was defined as bleeding requiring hospitalization and transfusion, geographic variations in the frequency of hospital admission or in transfusion policies may impact the apparent bleeding rates.

Conclusion

Important variations in outcome of stable atherothrombotic disease were observed according to broad geographic regions worldwide and persisted after multivariate adjustment. These observations have important implications for public health and clinical research.

Funding

This analysis was supported by Astra Zeneca and by the Département Hospitalo-Universitaire FIRE of Université Paris-Diderot and Assistance Publique – Hôpitaux de Paris.

Conflicts of interest

The authors would like to disclose the following.

Dr Ducrocq received consulting/speaking fees from Lilly and AstraZeneca.

Dr Deepak L Bhatt discloses the following relationships: Member of the Advisory Board of Medscape Cardiology; Board of Directors: Boston VA Research Institute, Society of Chest Pain Centers; Chair: American Heart Association Get With The Guidelines Science Subcommittee; honoraria: American College of Cardiology (Editor, Clinical Trials, Cardiosource), Duke Clinical Research Institute (clinical trial steering committees), Slack Publications (Chief Medical Editor, Cardiology Today Intervention), WebMD (CME steering committees); other: Senior Associate Editor, Journal of Invasive Cardiology; research grants: Amarin, AstraZeneca, Bristol-Myers Squibb, Eisai, Ethicon, Medtronic, Sanofi-aventis, The Medicines Company; non-funded research: FlowCo, PLx Pharma, Takeda.

R Gao received research grants from Sanofi-aventis, Medtronic, Boston Scientific, Abbott Vascular and MicroPort (Shanghai, China).

E Panchenko received speakers fees from Sanofi, Nycomed, Boehringer Ingelheim, Bayer, AstraZeneca, GlaxoSmithKline, The Medicines Company; honoraria for consultancy from Sanofi, Bayer, Lilly, AstraZeneca, Boehringer Ingelheim, Bristol-Myers-Sqibb, Pfizer, The Medicines Company; grants for clinical trials from Pfizer, Bristol-Myers Squibb, Boehringer Ingelheim, Sanofi, AstraZeneca, Daiichi Sankyo, Pharma Development, GlaxoSmithKline and DMPK.

Y Ikeda received grant support from Sanofi-aventis.

S Goto received research grants from Sanofi-aventis, Eisai, Otsuka, Boehringer Ingelheim; honoraria from Sanofi-aventis, Eisai and Otsuka; consulting fees from Merck and Eisai.

P Amarenco received research funding from Pfizer, AstraZeneca, Merck, Sanofi-aventis and BMS; speaker fees from Pfizer, Sanofi-aventis, Bayer, Boehringer Ingelheim, AstraZeneca and Otsuka; honoraria from Pfizer, Sanofi-aventis, Bristol-Myers Squibb, Merck, Kowa, Lundbeck, Bayer, and Boerhinger Inglheim.

PG Steg received a research grant (to INSERM U-698) from NYU School of Medicine, Sanofi, Servier; consulting/speaking fees from Amarin, AstraZeneca, Bayer, Boehringer Ingelheim, BMS, Daiichi Sankyo, GSK, Lilly, Medtronic, Otsuka, Pfizer, Roche, Sanofi, Servier, Takeda, The Medicines Company, Vivus; stockholding from Aterovax.

Acknowledgments

The REACH Registry was supported by Sanofi-aventis, Bristol-Myers Squibb and the Waksman Foundation (Tokyo, Japan). The REACH Registry is endorsed by the World Heart Federation.

Dr Steg and Dr Ducrocq had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

References