Design and Rationale of Gulf locals with Acute Coronary Syndrome Events (Gulf Coast) Registry

Registry Design

The Gulf COAST registry is a prospective, multinational, longitudinal, observational, cohort-based survey of all consecutive citizens who were admitted to hospitals with a confirmed diagnosis of ACS. Enrolment started in January 2012 and ended in January 2013. The Ethics Committees of each participating institution or country approved the study protocol.The current report presents the registry’s design, methodology, rationale, protocol implementation and patients’ baseline characteristics. The Gulf COAST registry adhered to the recommendations of “Strengthening the Reporting of observational studies in Epidemiology (STROBE)” [2].

Participating Hospitals

To provide findings that are representative of the involved countries, > 85% of the general hospitals in each country were included. The registry recruited patients from 29 hospitals in 4 Middle East Gulf countries: Kingdom of Bahrain (Bahrain): 2; State of Kuwait (Kuwait): 6; Sultanate of Oman (Oman): 12; and United Arab Emirates (UAE): 9.Of the 29 participating hospitals, 14 were community hospitals, 18 non-university teaching hospitals and 1 private sector hospital. All of the hospitals had on-site thrombolytic therapy available. On-site cardiac catheterization was available in 9 of the hospitals, while another 10 had a cardiac catheterization laboratory available to them within 1 h drive.

Administrative Organization and Hierarchy

Asteering committee from regional experts in the field was formed. A registry manager was appointed to coordinate between the national coordinators (NC), chief site officers (CSO) and site officers (SO) in the 4 countries. The protocol and the case report form (CRF) were drafted by the primary investigator and the study manager, and reviewed by the steering committee. Prior to the commencement of enrolment, one steering committee meeting and one investigators meeting were held. In addition, country-level and hospital-level meetings were held by the NCs in their respective countries. These meetings involved CSOs and SOs to educate them about the protocol and CRF. The CSOs supervised recruitment to ensure it was consecutive and helped answer data queries generated for their sites. The SOs recruited patients, and filled out CRFs prospectively at admission time and during hospital. Later on, they transferred the data online at their earliest convenience either from the workplace or other locations, depending on whether or not internet service was available at the hospital.

Study Population

Patients enrolled in the registry were citizens, 18 years of age or older with ACS diagnosed according to American College of Cardiology (ACC) clinical data standards [3].The decision to enroll citizens only was determined by the need to have near complete follow up of the study population.The expatriate population come from different countries, they are young, mostly male and highly mobile. This made it hard to guarantee their availability for follow up at 1 year. In addition, citizens are relatively homogenous in terms of their lifestyle, standard of living and access to free universal health care. There were no other exclusion criteria, and an attempt was made to recruit all comers. Patients were enrolled only after obtaining written informed consent to participate in the study. Patients who were eligible for inclusion in the registry but were not recruited for any reason (e.g. refused consent, died before consent) were entered into log sheets that recorded their age, gender, diagnosis and reason for not beingenrolled.

Data Collection and Validation

Data were prospectively collected on a standardized case report form (CRF). Data variables were in accordance with ACC key data elements and definitions for measuring the clinical management and outcomes of patients with ACS [3]. Data collected included patient demographics, previous diseases and risk factors, prior medication,clinical presentation, management during hospital stay including medications, reperfusion therapy and procedures, outcomes and discharge medications and follow up arrangements. Follow up was planned at 1, 6 and 12 months from the date of enrolment. Follow up was carried out by clinic visit or telephone interview. Data entry was carried out by SOs online at www.gulfcoastregistry.org.

A feature of the online data entry system included immediately-displayed automated data checks for out-of-range values and missing values of single data fields and also contradictory entries related to multiple data fields. In addition, throughout and at the conclusion of the study,to ensure accuracy of data, the registry manager carried out data cleaning using IBM® SPSS® Statistics 19 Program. A clinical research associate carried out site visits which included a visit to all participating hospitals at least once and involved random data source verification of 10% of paper and electronic CRFs at each visit.

Sample Size

In determining the sample size for Gulf COAST, we considered practical and statistical inputs. Practically, we weighed both logistical and funding issues including the number of citizens with ACS we expect to enroll from the participating centres over the planned enrolment phase.This was informed by prior experience with a large ACS registry of all comers in the Gulf Registry of Acute Coronary Events (Gulf RACE), where approximately 2500 local patients with ACS were enrolled over a 6-month period [4]. Hence, we can expect to comfortably enroll ~ 5,000 patients from the same centers over a 12-month enrolment period. Statistically, we recognize the difficulty in estimating sample size in an observational registry where several variables will be measured and no primary comparisons are planned. Nonetheless, we wanted to ensure adequate precision in the estimate of one-year all-cause mortality among locals presenting with ACS. To estimate an annual mortality rate of 10% at 1 year with a margin of error of 1% and 95% confidence intervals, a sample size of 2,948 patients is required. However, since there is little guidance in the literature on annual mortality of Gulf citizens with ACS, and also to compensate for losses of follow-up, we adjusted our sample size upwards to 4,000 patients.

As shown in Fig. (1), a sample size of 4,000 will provide adequate precision in estimating annual mortality over a reasonably wide range of event rates, and is logistically feasible given our prior experience [5]. We also considered the precision of all-cause mortality estimates for the following subgroups: ST-segment elevation myocardial infarction (STEMI), women, and patients younger than 50 years. Based on observations from Gulf RACE [5]. We estimate that 20% of citizens with ACS will have STEMI, 40% will be women, and 25% will be younger than 50 years. Therefore, for a total sample size of 4,000 patients with ACS, we expect 800 patients with STEMI, 1600 women, and 1000 patients younger than 50 years. These numbers will allow us adequate precision around all-cause mortality estimates for these subgroups with at least 90% confidence and, at most, 2% margin of error.

Statistical Analysis

Baseline hospital and patient characteristics were summarized using standard descriptive statistics.Continuous variables were summarized using means and standard deviations, where the distributions were normal or near normal, or medians and 25th/75th percentiles otherwise. Categorical variables were summarized using percentages.

Study Population Characteristics

Of 3188 patients, 1194 (37.5%) were recruited from Oman, 975 (30.5%) from Kuwait, 610 (19%) from UAE and 409 (13%) from Bahrain. The distribution of discharge diagnoses were as follows: non-ST-segment elevation myocardial infarction (NSTEMI): 1486 (46.6%), unstable angina: 931 (29.2%), STEMI: 741 (23.2%) and new-onset left bundle branch block myocardial infarction: 30 (0.9%). The mean age was 60.4 ± 12.8 years. Table 1 shows the patients’ characteristics.

The population risk profile included hypertension: 66%, dyslipidemia: 55.8%, diabetes mellitus: 53.4% and current smoking: 23.2% (Table 2).

At presentation, the mean heart rate was 85.6 ± 21.3 beats/min, mean systolic blood pressure was 141.3 ± 28.2 mmHg and mean diastolic blood pressure was 80.7 ± 16.5 mmHg. Three-quarters of the patients were over weight or obese (Table 3).

Non-Participation

A total of 151 patients fulfilled the entry criteria but were not enrolled in the registry because of a refusal to consent:

LBBB MI: left bundle branch block myocardial infarction; NSTEMI: non-ST-segment elevation myocardial infarction; STEMI: ST-segment elevation myocardial infarction; ACE: angiotensin converting enzyme; ARB: angiotensin receptor blocker; GPIIb/IIIa inhibitor: glycoprotein IIb/IIIa inhibitor.

TIA = transient ischemic attack; PCI = percutaneous coronary intervention; CABG = coronary artery bypass graft. 89 patients (59%) or not being approached by investigators: 62 patients (41%). Per country, the failure to enroll was 1.5% for Bahrain, 3.2% for Oman, 6.4% for Kuwait and 7% for UAE.

Implications of the Registry

We believe the subsequent analysis of all data and future publications from Gulf COAST will contribute to the existing literature and will contribute to better management of the ACS patient in the Gulf. We hope data from Gulf COAST would inform health care policy in the region and guide providers and regulatory authorities in their decisions regarding local and regional policy for cardiovascular care, from prevention to management and subsequent guideline construction. The data will provide local authorities with an insight into their local shortcomings which can translate into a quality improvement programs and policies.

Study Limitations and Strengths

As with any observational study, the execution of the registry may be limited by selection bias, particularly the potential exclusion of high-risk patients. In order to minimize this bias, we introduced a log sheet that investigators complete for all patients who were not enrolled. Moreover, we acknowledge that citizenship and ethnicity are not synonymous. Hence, while citizens may have relatively homogeneous lifestyle and social status, there may be some ethnic heterogeneity that we did not account for. The strength of our registry comes from its design that included 4 different countries, participation of the majority of hospitals in each country and enrolment of all patients presenting to each hospital.