Participants were identified through email and the Internet. Direct-to-patient emails with survey links were sent to two advocacy LISTSERVs maintained by professional organizations: Mended Hearts Inc., a cardiovascular patient membership organization, and the American College of Cardiology (ACC) Vendor “Opt-in” cardiovascular patient LISTSERV. In addition, a Web-based survey was available through an Internet link following search engine inquiries for “angina” or “angina treatment” and on the webpage of the Society for Cardiovascular Angiography and Interventions. [12] Unique US IP addresses were required. In both the email and the Internet link, a brief description of the survey was provided, and respondents opted in by clicking on the survey URL.

The target respondents were individuals aged ≥50 years who had a diagnosis of angina at any time in the past and also experienced angina in the prior 6 months. The survey was distributed during an 8-week period from November 2012 to January 2013. Respondents were invited to complete the survey if they met the eligibility by indicating yes to each criterion in the screening survey (age ≥50 years, provider diagnosis of angina, and occurrence of angina in the past 6 months). A description of angina was provided: “Angina is a pain, discomfort, or pressure localized in the chest that is caused by an insufficient supply of blood (ischemia) to the heart muscle. It is also sometimes characterized by a feeling of choking, suffocation, or crushing heaviness. This condition is called angina pectoris.” Respondents were excluded if they were employed or had an immediate family member employed by a marketing/research organization, or if they had participated in a research study on angina in the last 6 months. As an incentive, survey participants were offered entry into a drawing to win one of five $100 gift cards.

The survey was composed of 34 questions and took approximately 15 minutes to complete. The survey included demographic and clinical factors, contact with health care providers, medication use, and health-related QOL (HRQOL): overall health, angina burden, physical functioning, and social and emotional well-being. Angina and disease-specific HRQOL items were assessed using the Seattle Angina Questionnaire (SAQ). The SAQ is a 19-item self-administered questionnaire that measures 5 clinically important dimensions of coronary disease: physical limitation, angina stability, angina frequency, treatment satisfaction, and disease perception [13]. Domain scores range from 0 to 100, with lower scores indicating more angina and worse QOL. The specific measure of angina frequency used to stratify patients for analysis was based on the following question: “Over the past 4 weeks, on average, how many times have you had to take nitroglycerin for chest pain, chest tightness, or angina?” Patients were grouped according to the following responses: daily, ≥4 times daily, and 1-3 times daily; weekly, >3 times weekly but not daily, and 1-2 times per week; none/monthly, < once a week or no angina in last 4 weeks.

Global health status was assessed using the Global Health Question from the RAND Medical Outcomes Study (MOS) questionnaire, the EuroQol-5D questionnaire (EQ-5D), and the Work/Regular Physical Activity questionnaire. The 5-item MOS questionnaire is a self-assessment of a patient’s overall general health, with responses ranging from excellent to poor [14]. The EQ-5D is a 5-item tool that assesses a patient’s perception of his or her mobility, self-care, usual activities, pain/discomfort, and anxiety/depression [15]. EQ-5D responses can be converted to a standard scale using an algorithm developed for the US population [16], which range from 0 to 1, with 0 representing the worst imaginable health state and 1 representing perfect health [15]. The Work/Regular Physical Activity questionnaire is a 1-item questionnaire that assesses how active a subject is at work (including volunteer work and housework) [17]. The possible ordinal responses to the question are 1 = Mainly sedentary, 2 = Predominantly walking on one level, 3 = No heavy lifting, 4 = Mainly walking, including climbing stairs, walking uphill, or lifting heavy objects, 5 = Heavy physical labor, and 6 = Do not work. The response “Do not work” is not applicable to the physical activity at work endpoint, and therefore, individuals with this response were excluded from analysis of this domain of health.

Respondents confirmed that they had received a provider diagnosis of angina. However, the main analysis was further limited to respondents who also reported a confirmatory CHD history—who reported taking antiplatelet agents and had a prior percutaneous coronary intervention (PCI), prior coronary artery bypass grafting (CABG), or coronary heart disease diagnosis. Demographics, clinical history, comorbidities, general clinical care, and HRQOL were compared according to SAQ angina frequency groups. Supplementary analyses for the non-CHD confirmed cohort are shown in Multimedia Appendix 1, Multimedia Appendix 2, and Multimedia Appendix 3. The purpose of the analysis was descriptive and no specific hypotheses were prespecified. Analyses were performed using SAS Version 9.4.

All pages of the survey required complete answers prior to advancing; therefore, there were no missing items. As participants voluntarily submitted responses to this anonymous survey, the Duke institutional review board waived the need for consent. The study concept originated with the sponsor, Gilead Sciences. The sponsor contracted with members of the Outcomes Group at Duke Clinical Research Institute (DCRI) to design the survey, perform the analyses, and report the results. The DCRI retained full independence in matters of analysis, interpretation, and publication. The sponsor was given an opportunity to review and comment on the publication prior to submission, but final responsibility for content remained with DCRI authors. One sponsor representative who worked with DCRI on the study design was included as a coauthor.

This anonymous angina Internet survey provides insights into the range of symptomatic individuals in the community and raises questions about their identification and treatment. We found that approximately 15% of respondents reported daily angina, and approximately 40% reported weekly angina. In both the CHD group and non-CHD group, more frequent angina was associated with worse QOL. The cardiac origin of the symptoms cannot be proven with the methods we used; however, similarities in comorbidity, treatments, and symptoms in both the CHD and non-CHD group suggest that the burden of angina is similar in both. The majority of respondents had seen a cardiologist in the prior 6 months, yet medication and revascularization seemed lower than expected given their symptomatic status. While this methodology was successful in rapid accrual of data directly from patients, responses must be viewed in context. The Internet may be able to build a bridge for symptomatic individuals to enter the clinical setting for more timely recognition and treatment of angina.

Angina in population surveys, trials, and registries also takes place through patient report without objective confirmation, which elevates the comparability of our findings [18,19]. Most of the reported angina is likely from coronary artery disease as a majority had a prior revascularization. Angina typically persists in approximately 20% of post-MI or prior revascularization patients and exists in the same proportion among a population with stable ischemic heart disease [20]. However, up to 40% of angina patients, even those with a history of CHD, who undergo angiography have nonobstructive coronary artery disease [21]. Angina in the absence of epicardial disease may come from microvascular dysfunction, spasm, and diffuse plaques [22,23]. The possible existence of angina without ischemia or epicardial disease is important, as some providers may discount symptoms in the absence of obstructive coronary disease. For respondents in this survey with a history of angina, daily angina occurred in ~15%, and weekly angina in another 40%. Some of the subjects may have nonanginal chest pain, or somatic manifestations of depression; however, the data suggest the cause is believed to be angina from the patient perspective. Therefore, to succeed in alleviating “angina” as a source of patient suffering and impaired QOL, a broader perspective on its identifying criteria seems warranted. Requiring the presence of significant angiographic disease or abnormal stress testing may be overemphasized. Angina should be an important criterion for study inclusion in its own right.

Respondents with more angina were younger, more often female, less likely to have had prior PCI or CABG, on fewer antianginal medications, and more depressed (34.1 vs. 19.6%). Only 27% of those with daily angina reported freedom from anxiety or depression on the EQ-5D survey. Comparing to similar populations from trials and registries is challenged by the inclusion of patients with other qualifying criteria such as angiographic disease or ischemia, in addition to some degree of angina. The TERISA population enrolled patients with chronic angina in addition to diabetes [24]. The use of long-acting nitrates in the TERISA population was similar to the respondents with daily or weekly angina in this survey, with most other preventive medication use being similar. The CLARIFY registry population enrolled patients with stable ischemic heart disease, of whom 20% had angina with or without documented ischemia. In the CLARIFY population, those with angina were younger, female, and had less frequent history of revascularization than those without angina [20]. We observed that antianginal medication use increased among those with more reported angina, yet only 44% with daily symptoms were on two or more of these medications [6]. Control of blood pressure, cholesterol, and weight may also be low in this group. Of those reporting daily or weekly angina, 10-15% did not discuss it with their cardiologist, making effective treatment unlikely. Angina is associated with a significant increase in CV-related death or MI across every Canadian Cardiovascular Society Angina class [20]. This association is probably due primarily to the effect of atherosclerotic coronary artery disease; however, linkages with other prognostically important disorders such as depression might also contribute. This Internet population was identified purely through patient-reported symptoms, adding a unique comparator to symptomatic populations assembled using other inclusion criteria, and underscoring that those with chronic angina are often female and younger.

It is possible that the qualifying angina in this survey population occurred between visits to providers. The identification of a population with angina (via the Internet) raises the question of access in the event of symptom return between clinic visits. Strategies such as the brief SAQ instrument for patients with coronary disease or angina as part of routine visits, or between visits, could identify those likely to benefit from treatment intensification or revascularization [25]. Despite screening, variation in the level of angina control across health care clinics is also known to exist [26]. Although the vast majority reported having health insurance, 15% of the daily angina group also reported having no cardiologist. This suggests a lost opportunity in access to health care. Regional clinic accessibility, travel limitations, or social barriers to receiving clinic-based care may be contributors in this population.

To participate in an Internet survey, respondents must have computer access, which may limit generalizability [27]. As responses were voluntary, individuals with a greater burden of angina may have been more motivated to participate, leading to an overestimation of the effect of angina on QOL. Demographic or clinical data on nonrespondents could not be assessed. In addition, data on clinical diagnoses and medications were based on self-report. However, profiles and angina severity are consistent with those of contemporary populations with angina in more traditional clinical studies [28]. The possibility remains that a diagnosis of angina, as required by the screener, was misunderstood. We selected the CHD cohort for the main analysis, limited to those with documented coronary disease, to address this concern. Lastly, there was no mechanism to verify unique responses; however, each IP address could only be associated with one completed survey. This study describes responses at a single point in time, so follow-up information on treatment or outcomes was not available.

This survey provides a snapshot of those with angina in an online community—at a single time point, across providers and treatment stages—and finds 15% of this group experiences daily angina. This study suggests the promise of Internet surveys for assessing patient-reported symptoms, and raises the possibility that screening tools such as this could be deployed inexpensively between health care provider encounters for better angina control and potentially improved QOL.