Protocol of the RED-CVD study
Background
The diagnosis of cardiovascular diseases (CVDs) in the early stages is notoriously difficult, as the first symptoms are often non-specific or atypical. It is even more challenging in patients with comorbidities such as COPD or T2D, because of similarities in signs and symptoms, such as dyspnea or reduced exercise tolerance. Screening studies in patients with COPD or T2D, who are at highly increased risk for developing CVD,1–4 showed that unrecognized heart failure was present in 20% of patients with COPD and 28% of patients with T2D.5,6 Unrecognized coronary artery disease was even more prevalent; in 65% of patients with T2D.7
In the Netherlands, the majority of patients with COPD or T2D are managed in primary care disease management programs. Even though it is well-established that CVDs contribute significantly to both morbidity and mortality in COPD8,9 and T2D,10–12 no attention is focused on the recognition of new-onset CVD within these programs. A pro-active approach, focusing on the detection of CVD in these elevated-risk individuals could lead to more intensive medical management and subsequently reduce adverse cardiovascular morbid events. This is of particular importance for CVDs with evidence-based therapies known to improve prognosis, such as coronary artery disease (CAD), atrial fibrillation (AF) and heart failure (HF).
We developed a simple, two-stage diagnostic strategy for uncovering the target CVDs, starting with a questionnaire on signs and symptoms suggestive of CVD, which is followed by electrocardiography, physical examination and NT-proBNP measurement in patients who indicated having symptoms. This Early Diagnosis Strategy is to be blended in with the existing primary care disease management programs for COPD and T2D. In the current study, we will investigate the yield of this strategy compared to usual care, in terms of newly recognized CVD and subsequent targeted treatment. We will also assess whether the addition of family history and, in females, reproductive history taking have added diagnostic value. Lastly, we will investigate the diagnostic value of several cardiac biomarkers, i.e. high sensitive troponin I (hs-Tn I) , ST-2 and growth differentiation factor 15 (GDF-15), for the detection of the target CVDs.
Key objectives
To determine the diagnostic yield of the ‘early diagnosis strategy’ compared to usual care, in terms of detection of previously unrecognized CAD, AF and HF in patients with COPD or T2D.
To assess the added diagnostic value of family- and female-specific history taking.
To assess the added diagnostic value of as three cardiac biomarkers: high sensitive troponine I, growth differentiation factor 15 and ST-2
To assess the effect of the Early Diagnosis Strategy and subsequent targeted interventions on health-related quality of life.
To determine the cost-effectiveness of the Early Diagnostic Strategy.
Methods
Study design
A cluster randomized trial will be conducted, with general practices at the unit of randomization, comparing the effects of the Early Diagnosis Strategy to usual care in terms of newly recognized CAD, AF and HF. Additionally, in a diagnostic study in the intervention arm, the added diagnostic value of the family- and female-specific questionnaires as well as three cardiac biomarkers will be evaluated.
Study population and recruitment
Patients who participate in a primary care disease management program for COPD or T2D are selected from the electronic medical files of the participating general practitioners (GPs) in the central and northern regions of the Netherlands. Of all patients with T2D in the Netherlands, 84% participates in the primary care disease management program for T2D. The participation rate of primary care disease management program for COPD is lower, approximately 65%, as patients with GOLD III-IV stage COPD are under specialist pulmonary care. Our sample can therefore be considered as a representative sample of patients with T2D or GOLD I-II stage COPD.
Patients with an established triple diagnosis of CAD, HF and AF (confirmed with echocardiography in case of HF, with coronary angiography, stress tests or a calcium score > 100 on CT-scan in case of CAD, and with electrocardiography in case of AF) are excluded. Patients with one or two established diagnoses are not excluded, because they are at an increased risk for having a second or third CVD, that could potentially be uncovered by the Early Diagnosis Strategy.
The participating GP sends a letter to all eligible patients; patients who are interested in participation will receive further information and are asked for informed consent by the researcher before any study procedure is undertaken.
Randomization
General practices are randomly allocated to either the Early Diagnostic Strategy or care as usual. Randomisation is necessary because we aim to quantify the clinical effectiveness or yield of the Early Diagnostic Strategy. We will use cluster randomisation instead of randomisation at an individual participant level to avoid contamination of participants and GPs between the groups. A computer-based randomisation will be performed with use of the minimization method based on general practice size, to avoid imbalances in numbers of participants in both arms of the trial.
Sample size
Based on the prevalence of unrecognized HF, CAD and AF detected in patients with COPD or T2D in previous screening studies, we estimate to uncover these target CVDs in 10% of the patients with the Early Diagnosis Strategy, compared to 5% with usual care. To detect this 5% difference in new diagnoses, using an alpha of 0.05, a power of 0.80, an intracluster correlation coefficient of 0.01 and an anticipated drop-out of 10%, will require 650 participants per arm. To recruit 1300 patients, around 40 general practices should participate.
Study procedures
The Early Diagnosis Strategy will be blended in with the existing primary care disease management programs. No additional visits are necessary as all study procedures will take place during the routine appointments that are part of the programs. Patients in both arms will fill out a questionnaire on general health-related quality of life, the EQ-5D-5L, before their next routine visit. During the routine visit, patients in the intervention arm will undergo the Early Diagnosis Strategy and patients in the control arm will receive care as usual. After one year follow-up, medical files of all patients are searched to document new diagnoses, changes in medication status and use of care during the past year, and patients are asked again to fill out the EQ-5D-5L.
The Early Diagnosis Strategy and care as usual
The Early Diagnosis Strategy is a two-stage intervention. The first step consists of a questionnaire on signs or symptoms suggestive of CVD. This ‘Early Diagnosis Questionnaire’ will be filled out at home, prior to the routine visit that was scheduled for the COPD- or T2D primary care disease management program. The GP will receive the answers. All patients who score 1 point or higher on the questionnaire, indicating they have symptoms suggestive of CVD, will proceed to the second stage of the intervention, which consists of a focused physical examination, electrocardiography (ECG) and NT-proBNP measurement.
The physical examination includes an inspection of legs and neck veins, palpation of the radial pulse and apex, and auscultation of heart and lungs. Physical examination will be performed by a trained practice nurse and general practitioner in a standardized manner. Signs of fluid overload (peripheral oedema, pulmonary crackles, increased jugular venous pressure), irregular pulse or tachycardia, abnormal heart sounds or murmurs, or a laterally displaced apex beat will be considered abnormal. A standard 12-lead ECG is recorded by a trained employee of the GP or GP’s laboratory and will be interpreted by the GP. Abnormalities include rhythm disturbances, left axis deviation, conduction abnormalities, pathological Q-waves, voltage criteria for left ventricular hypertrophy, T-wave abnormalities and ST-deviations. Venous blood samples are taken and B-type Natriuretic Peptide (NT-proBNP) levels will be determined with a pre-specified cut-off point of 125 pg/mL.
In case of one or more abnormal findings in the second stage of the intervention, the GP is recommended to consult a cardiologist or refer the patient for further specialist investigations. Patients in the control arm will receive care as usual in the COPD- or T2D disease management programs.
Diagnostic study in the intervention arm: added value of family and reproductive history and three cardiac biomarkers
Patients in the intervention arm will be asked to fill out additional questionnaires on family history of CVDs and, in women, on reproductive history and pregnancy complications. The questionnaires will be provided with the Early Diagnosis Questionnaire, prior to the patient’s routine visit, but will not be part of the Early Diagnosis Strategy and analyzed in a separate diagnostic study.
In addition, from all patients in the intervention arm who undergo blood sampling for NT-proBNP measurement, an extra 10 mL blood sample will be drawn during the same venapuncture. After centrifugation, specimens of serum will be stored at -80 degrees Celsius. At the end of the follow-up period, serum samples will be thawed and levels of GDF-15, ST-2 and hs-Tn I will be determined. Remaining samples are stored in a biobank to enable future biomarker studies.
Outcome measures
Heart failure, atrial fibrillation and coronary artery disease
Presence of the target CVDs will be determined one year after the initial investigations. The final diagnoses are set by a joint panel consisting of three members (either two cardiologists and one general practitioner or three cardiologists), who will evaluate the results of diagnostic tests that took place during the follow-up period. Adjudication of diagnoses by a panel of physicians is an accepted reference standard for the evaluation of heart failure, for which a true criterion standard is lacking. The presence of appropriate symptoms and signs as well as an appropriate measure of cardiac dysfunction will be evaluated by the panel, according to the guidelines of the European Society Cardiology (ESC). Heart failure is considered present when the participants mention suggestive symptoms in the Early Diagnosis Questionnaire (e.g. exertional dyspnea, orthopnea, peripheral edema or nycturia) or have suggestive signs at physical examination (e.g. laterally displaced apex, pulmonary rales, peripheral oedema and/or raised jugular venous pressure), in combination with echocardiographic evidence of cardiac dysfunction at rest. In participants who are already on treatment with diuretics, signs of volume overload could be masked and are therefore not obligatory.
For coronary artery disease and atrial fibrillation, widely accepted criterion standards exist. For the diagnosis of CAD, a positive stress test (e.g. ST-deviation on an exercise test or a positive (dobutamine) stress echocardiography, stress SPECT, stress MRI or stress PET) or any imaging modality showing anatomical proof for coronary artery disease (coronary angiography, coronary CT-A or a calcium score > 100 with CT-scan) is prerequisite.13 For the diagnosis of atrial fibrillation, rhythm documentation is required using a standard 12-lead ECG (or a single lead-ECG of at least one minute) showing the typical pattern of AF: absolutely irregular RR intervals and no discernible, distinct P waves.
In case consensus is not reached, a majority decision will be used. There will be a random sample re-testing of 10% to evaluate the reproducibility of the panel diagnoses.
Health-related quality of life
Health-related quality of life is measured with the EQ-5D-5L questionnaire. The EQ-5D-5L comprises five dimensions (mobility, self-care, usual activities, pain/discomfort and anxiety/depression) which are divided into five degrees of severity, ranging from ‘no problem’ to ‘extreme problems’. In addition, the patient’s self-rated health is recorded on a visual analogue scale (VAS). The EQ-5D-5L produces a single index score.
Costs
Information on diagnostic investigations, outpatient visits, hospitalizations and prescribed medication will be collected from the GP’s electronic medical files for patients in both arms. Additionally, costs for performing the Early Diagnosis Strategy will be calculated for the intervention group. Prices will be based on market prices and tariffs.
Biomarkers
At the end of the study, the added diagnostic value of GDF-15, Hs-Tn I and ST-2 will be evaluated for the clinical endpoint of presence or absence of previously undetected HF, AF or CAD.
Family history and reproductive history
The added diagnostic value of a positive family history for CVDs in first or second degree relatives will be assessed. In women, the use of detailed information on reproductive history (early menopause, gestational hypertension or (pre-)eclampsia, gestational diabetes, miscarriages or preterm birth) will also be evaluated.
Blinding
The study will be conducted in a single blind manner. GPs are invited to participate in a study investigating the early detection of CVD in high-risk patients with COPD or T2D and will receive more detailed information after their practice has been randomized. Selectively informing the GPs is to prevent that GPs in the control group will apply (components of) the Early Diagnosis Strategy. GPs and participants in the control group will receive information on study procedures performed in the intervention group after the study has been completed. Researchers are not blinded. Since the questionnaires are filled out at home prior to routine visits, researchers are not at risk for observation bias. The panel will be blinded for allocation to intervention or usual care.
Statistical analyses
The yield in terms of previously unrecognized CVD will be calculated as a proportion with 95% confidence intervals in both study arms, as well as the absolute difference in yield between the arms. Given that this is a cluster-randomized study, we anticipate that (few) differences in baseline risk will be present in both groups. Hence, we will perform both an unadjusted and adjusted analysis and assess whether the effect of the Early Diagnosis Strategy changes. The difference in health-related quality of life between the intervention and control arm at the one year follow-up mark will be compared, again taking into account potential baseline differences.
In the intervention arm, detailed family history and (in women) reproductive history will be considered a diagnostic variable and the added value to the Early Diagnosis Strategy will be evaluated with multivariable logistic regression analysis, for the outcome of presence or absence of CAD, AF, and HF.
The added diagnostic information from three biomarkers (GDF-15, hs-Tn and ST-2) will be quantified by using multivariable logistic regression analysis, with CVD presence as the binary outcome of the model. Biomarkers will be added as continuous variables to a model including the results of the Early Diagnosis Strategy, using interaction terms for other biomarker results, gender and age.
Cost-effectiveness analyses
The cost-effectiveness of the Early Diagnosis Strategy will be evaluated in comparison with care as usual. Cost-effectiveness is expressed in terms of costs (or cost savings) per additional correct diagnosis and in terms of costs per additional Quality Adjusted Life Year (QALY) gained. For measuring direct costs, data on resource use will be collected, based on observed volumes of each of the included tests, the number of consultations (with cardiologists) and additional investigations, the volumes of preventive medication and treatment prescribed. A decision analytic model will be developed for long term consequences, for which actual CVD risks from the current study will be complemented with evidence on treatment effects, CVD risks and long term consequences of CVD events gathered from literature. Comprehensive sensitivity analysis will be performed to determine the robustness of the cost-effectiveness results. Results will be presented in incremental cost-effectiveness ratio (ICER) planes and as cost-effectiveness acceptability curves, separately for men and women.
Regulation statement and ethics committee approval
This study is conducted according to the principles of the current version of the declaration of Helsinki and in accordance with the Dutch law on Medical Research Involving Human Subjects Act.
The study was approved by the medical ethical committee of the University Medical Center Utrecht.
Results of inclusion
?
Discussion
This trial will generate evidence on the diagnostic yield of the Early Diagnosis Strategy in terms of detection of previously unrecognized heart failure, coronary artery disease or atrial fibrillation in elevated-risk individuals with COPD or T2D. The Early Diagnosis Strategy will be implemented in the existing primary care disease management programs for COPD and T2D. Comparison with usual care will reveal whether the Early Diagnosis Strategy and subsequent targeted therapies are beneficial for this population, with regard to newly detected cases as well as health-related quality of life. Cost-effectiveness will be investigated.
Relevance and clinical applicability
If convincingly effective, the Early Diagnosis Strategy is suitable for immediate implementation in disease management programs for T2D and COPD. The intervention aims to improve the detection of early stage CVD by focusing on symptoms, and evaluating these symptoms with objective measures. It can therefore be easily applied to other elevated-risk populations as well, as symptoms are expected to be similar and ECG, NT-proBNP and physical examination have been well-established and validated diagnostics for a range of patient populations.
Relation to other studies
It is now well established that CVDs contribute significantly to both morbidity and mortality in COPD and T2D. Even so, screening studies reported that unrecognized, new-onset CVDs are still highly prevalent in these high-risk individuals.2,5,14–17 In many of these studies, a large proportion of patients did indicate having symptoms when asked proactively, suggesting that the early phases of CVD might often be symptomatic but overlooked, rather than asymptomatic.2,14,15 Especially in patients with comorbidities, early signs and symptoms of CVDs are easily missed; either because they are so non-specific that patients do not mention them to their GP, or because they are so similar to symptoms of the established comorbid disease that doctors do not recognize them as being first signs of a new, cardiovascular pathology.
Nonetheless, whether screening for CVDs is useful remains controversial, even for elevated-risk populations. The recently published, updated US Preventive Services Task Force (USPSTF) recommendations, cite a lack of evidence to assess the net benefits vs net harms of screening ECG in high-risk patients.18 The American College of Cardiology recommends against screening in asymptomatic high risk patients owing to a lack of outcome benefit, and emphasizes the potential risk of procedural harms, as abnormal screening results often require additional (invasive) investigations.19
The strategy proposed in the current study, contrary to screening strategies, primarily focuses on uncovering symptoms. The incremental value of implementing the Early Diagnosis Strategy in the existing disease management programs will therefore not be a reclassification of CVD risk, but a chance to perform a precise and targeted therapeutic intervention in symptomatic patients, to improve quality of life by reducing symptom burden and preventing progression into morbid CVD events.