Objectives Coronary angiography is regularly performed in patients with worsening signs and/or symptoms of heart failure (HF). However, little is known on the determinants, findings and associated clinical outcomes of coronary angiography performed in patients with worsening HF.
Methods The BIOSTAT-CHF (a systems BIOlogy Study to TAilored Treatment in Chronic Heart Failure) programme enrolled 2516 patients with worsening symptoms and/or signs of HF, either hospitalised or in the outpatient setting. All patients were included in the present analysis.
Results Of the 2516 patients included, 315 (12.5%) underwent coronary angiography within the 30 days after the onset of worsening symptoms and/or signs of HF. Subjects who underwent angiography were more often observed as inpatients, had more often an overt acute coronary syndrome, had higher troponin I levels, were younger and had better renal function (all p≤0.01). Patients who underwent coronary angiography had a lower risk of the primary outcome of death and/or HF hospitalisation (adjusted HR=0.71, 95% CI 0.57 to 0.89, p=0.003) and death (adjusted HR=0.59, 95% CI 0.43 to 0.80, p=0.001). Among the patients who underwent coronary angiography, those with a coronary stenosis (39%) had a worse prognosis than those without stenosis (adjusted HR for the primary outcome=1.71, 95% CI 1.10 to 2.64, p=0.016).
Conclusions Coronary angiography was performed in <13% of patients with symptoms and/or signs of worsening HF. These patients were remarkably different from those who did not undergo coronary angiography and had a lower risk of subsequent events. The presence of coronary stenosis on coronary angiography was associated with a worse prognosis.
- decompensated heart failure
- coronary angiography
- acute coronary syndrome
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Coronary angiography is the ‘reference standard’ technique for the assessment of the presence and the extent/severity of coronary artery disease, and to define the most appropriate therapy.1 Current heart failure (HF) guidelines state that coronary angiography is recommended for the determination of HF aetiology, especially in patients who suffer from angina pectoris, those with a history of ventricular arrhythmia or aborted cardiac arrest, and in patients with intermediate to high pretest probability of coronary artery disease, which includes a ‘positive’ non-invasive stress test.2 3
In patients with worsening symptoms and/or signs of HF, coronary angiography may be infrequently performed, regardless of hospitalisation or ambulatory status.4 5 However, little is known about the type of patients who undergo coronary angiography, whether significant coronary artery disease is found, and whether it has prognostic implications.
The aims of the present analysis are to assess: (1) related factors and characteristics of patients with worsening HF who undergo coronary angiography; (2) the findings of coronary angiography regarding the presence of coronary stenosis; (3) the prognostic value of coronary angiography and coronary stenosis.
BIOSTAT-CHF (a systems BIOlogy Study to TAilored Treatment in Chronic Heart Failure) is a European project that enrolled 2516 patients with HF from 69 centres in 11 European countries to determine profiles of patients with HF who do not respond to recommended therapies, despite anticipated uptitration. The design and first results of the study and patients have been described elsewhere.6 In brief, patients were aged ≥18 years with symptoms of new-onset or worsening HF, confirmed either by a left ventricular ejection fraction (LVEF) of ≤40% or a B-type natriuretic peptide (BNP) and/or N-terminal proBNP (NT-proBNP) plasma levels >400 pg/mL or >2000 pg/mL, respectively. Patients needed to be treated with either oral or intravenous furosemide ≥40 mg/day or equivalent at the time of inclusion. Patients should not have been previously treated with evidence-based therapies (ACEi/ARBs (ACE inhibitor/angiotensin receptor blockers) and β-blockers) or were receiving <50% of the target doses of at least one of these drugs at the time of inclusion. Initiation or uptitration of ACEi/ARB and/or β-blocker therapy should have been anticipated by the treating physician. The first 3 months of treatment were considered to be the optimisation phase after which a stabilisation phase of 6 months was defined. During the optimisation phase, initiation or uptitration of ACEi/ARB and/or β-blocker was performed according to the routine clinical practice of the treating physicians, who were encouraged to follow the European Society of Cardiology guidelines at the time of treatment.7 8 Patients with acute coronary syndrome (ACS) or stroke could be included when the primary diagnosis for admission to hospital or outpatient clinic visit was HF.6 The recruitment period was 24 months, starting from December 2010. The last patient was included on 15 December 2012. Median follow-up was 21 months.
In this post hoc analysis, we included all coronary angiographies performed within 30 days after the baseline visit, because coronary angiography could have been done as ‘programmed intervention’ and, therefore, a time gap between the intervention and the baseline visit was expected. Coronary stenosis was defined as >50% luminal stenosis (online supplementary table 1). Participating European countries were also divided into North and South for comparison as follows: Northern Europe—Netherlands, Norway, Sweden, Germany and UK; Southern Europe—France, Slovenia, Italy, Greece, Serbia and Poland. A subanalysis by country was also performed.
Supplementary file 1
Ethics board approval was obtained and all participants signed written informed consent before entering the study.
In descriptive analyses, continuous variables are expressed as mean±SD. Categorical variables are expressed as frequencies and proportions (%). Population description and comparison of patients with coronary angiography versus no coronary angiography performed (and coronary artery coronary stenosis vs no stenosis) was performed using independent samples t-test for normally distributed continuous variables, Mann-Whitney U test for continuous variables with a skewed distribution and χ2 test for categorical variables. Normality assumptions were verified by visual inspection. No multiple imputation was performed.
To determine the factors associated with having coronary angiography performed (or not) and to having a coronary artery coronary stenosis (or not), we developed logistic regression models. These models used clinical and laboratory variables with a p value <0.1 as entry criteria (from table 1). Logistic regression assumptions were checked and multicollinearity excluded. Linear relationship between continuous independent variables and the logit transformation of the dependent variable was verified by plotting the means versus the β estimates in quintiles (online supplementary figure 1). If a linear relationship was not present, then the variable was dichotomised at the inflexion point. Then a stepwise backward selection process was applied and the final model presented.
Cox proportional hazard regression models were used to model long-term event rate both in univariable and multivariable analyses. Proportional hazard assumption was verified graphically using ‘log-log’ plots. In the multivariable models, the covariates for adjustment were chosen from demographic (age and gender), clinical (previous HF hospitalisation, use of β-blockers and systolic blood pressure) and laboratory (NT-proBNP, blood urea nitrogen, haemoglobin, high-density lipoprotein-cholesterol, creatinine, sodium). All parameters were previously found to be independently associated with the outcomes in the BIOSTAT cohort and were used to build the risk models derived from this cohort (URL: https://biostat-chf.shinyapps.io/calc/).9
The primary outcome was a composite of hospitalisation for HF and all-cause death. The outcomes of HF hospitalisation and death were also analysed separately. For the outcome of HF hospitalisation, a competing risk model (using death as competing risk) was used according to the method of Fine and Gray.10
The adjudication of events (HF hospitalisations) was done by the treating physician.
All the analyses were performed using R software (R Core Team, 2013. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL: http://www.R-project.org/). The competing risk models and proportional hazard assumption were performed using STATA (StataCorp. 2015. Stata Statistical Software: Release 14. College Station, TX: StataCorp LP.).
Characteristic of the study population
From the 2516 patients included in BIOSTAT-CHF, 12.5% (n=315) underwent coronary angiography.
Characteristics of patients with or without coronary angiography are presented in table 1. Patients who underwent coronary angiography more often presented as inpatients, with an ACS, were younger, had higher heart rate, haemoglobin, estimated glomerular filtration rate (eGFR), alanine/aspartate aminotransferase and troponin I levels. The troponin I threshold for coronary angiography performance was high: only patients in the highest troponin quintile (>36 pg/mL) were more likely to have a coronary angiogram performed (online supplementary figure 1). Nonetheless, troponin I levels were linear and independently associated with worse prognosis in this population but added little prognostic information to the BIOSTAT risk models (online supplementary tables 2 and 3). Patients who underwent coronary angiography were also more often smokers and more frequently treated with ACEi/ARBs, had lower LVEF, urea and potassium, were less often hospitalised in the year before baseline visit, had ischaemic cardiomyopathy less often documented, had lower proportion of atrial fibrillation, previous stroke, device therapy and previous coronary intervention (p<0.01 for all) (table 1). Country subanalysis shows that the Netherlands and France had the higher proportion of patients undergoing coronary angiography. Of notice, the Netherlands contributed with more than 25% of all angiographies performed (online supplementary table 4).
Independent predictors for performing coronary angiography are presented in table 2. The strongest independent predictors of undergoing coronary angiography were an in-hospital visit (OR=11.6, 95% CI 4.6 to 28.8, p<0.0001), overt ACS (OR=3.1, 95% CI 1.9 to 5.0, p<0.0001), troponin I levels above 36 pg/mL (OR=1.6, 95% CI 1.1 to 2.3, p=0.011), a younger age (OR per each decade less=1.4, 95% CI 1.2 to 1.6, p<0.0001) and better renal function (OR per 10 mL/min/1.73 m2 increase in eGFR=1.1, 95% CI 1.0 to 1.2, p=0.049). Patients with a cardiac device, those with previous HF hospitalisation and those with previous coronary intervention were less likely to have coronary angiography performed (table 2).
Coronary angiographic findings
A coronary stenosis (>50% luminal stenosis) was found in 38.7% (n=122) of the 315 patients who underwent coronary angiography. Characteristics of patients with and without a coronary stenosis are presented in the online supplementary table 1. Patients with a coronary stenosis were older, more often male, smokers and hypertensive, had higher proportion of pulmonary rales, HF of ischaemic aetiology more often documented, higher troponin I levels and higher proportion of previous coronary intervention (p<0.01 for all).
Among the patients who underwent coronary angiography, those with HF of ischaemic aetiology (OR=33.4, 95% CI 16.4 to 68.0, p<0.0001) and with higher troponin I levels (OR per 1 log increase=1.3, 95% CI 1.0 to 1.7, p=0.026) were more likely to have a coronary stenosis (table 3).
Prognostic implications of coronary angiography and presence of coronary stenosis
Patients who underwent coronary angiography had a better clinical outcome compared with those who did not undergo coronary angiography (adjusted HR, HR for the primary composite outcome of death and/or HF hospitalisation=0.71, 95% CI 0.57 to 0.89, p=0.003 and HR=0.59, 95% CI 0.43 to 0.80, p=0.001 for the outcome of death) (table 4). Among the patients who underwent coronary angiography, those with a coronary stenosis had worse prognosis (adjusted HR for the primary composite outcome of death and/or HF hospitalisation=1.71, 95% CI 1.10 to 2.64, p=0.016 and HR=2.09, 95% CI 1.10 to 3.96, p=0.024 for the outcome of death) (table 4).
A significant interaction between HF aetiology (ischaemic vs other) and coronary angiography (yes vs no) was found. Patients who underwent coronary angiography with non-ischaemic HF had a greater reduction of the primary composite outcome (HR=0.55, 95% CI 0.40 to 0.76, p<0.001) than patients who underwent coronary angiography with ischaemic HF (HR=1.00, 95% CI 0.74 to 1.37, p=0.98; p for interaction=0.007) (figures 1 and 2).
The present study shows that ≈13% of patients with worsening HF underwent coronary angiography within 30 days of presentation for worsening symptoms and/or signs of HF. In general, these patients had a better clinical profile and outcome than those who did not undergo coronary angiography. However, patients with a coronary stenosis on coronary angiography had a worse prognosis compared with those without a coronary stenosis.
In our study, the coronary angiography rate was higher than in previous reports where less than 10% of the patients with worsening HF underwent coronary angiography.5 11 Nonetheless, in patients with decompensated HF, coronary artery disease may be the primary HF aetiology in more than 50% of the patients.12 Hence, addressing coronary artery disease as a therapeutic target in worsening HF (even without overt ACS) may be associated with improved clinical outcome, and although a causal relation cannot be inferred, recurrent ischaemic events are a major cause of subsequent HF decompensation and death.13
In the present report, only 23% (n=54) of the subjects presenting with an overt ACS underwent coronary angiography within the worsening HF episode (±30 days). These data suggest that the large majority of the coronary angiographies were performed in patients with other primary causes for HF decompensation. Hence, in the present study, physicians possibly decided to perform coronary angiography based on the suspicion that an underlying coronary artery disease was a major contributor for worsening HF signs and/or symptoms (also supported by particularly high troponin threshold for angiography performance). Troponin elevation is frequently observed in patients with decompensated HF, possibly reflecting myocardial injury and/or impaired myocardial perfusion, and has been associated with worse prognosis.14 While doctors acknowledge troponin elevation as part of the decompensation episode, they may withhold coronary angiography unless very high troponin levels are found, because despite the myocardial injury, patients with decompensated HF may have a predominance of respiratory symptoms, high prevalence of diabetes and use of medications such as nitrates, β-blockers and ivabradine that may blunt ‘typical’ angina pectoris symptoms.15 16 Hence, from a clinical standpoint it may be challenging to distinguish between a primary diagnosis of ACS with associated HF versus worsening HF with elevated troponin without ‘typical’ symptoms of ischaemia. In selected cases, tests for myocardial viability and ischaemia and/or coronary angiography may help make therapeutic decisions.
A large country variability in the performance of coronary angiography was also found, notably more than 25% of the angiographies were performed in the Netherlands and more than 15% in France. These findings may reflect country variation in the accessibility to a catheterisation laboratory.
Diagnostic procedures may influence treatment decisions (directly and/or indirectly) and consequently prognosis.17–19 In this context, the performance of coronary angiography may provide information regarding the extent/severity of coronary artery disease and also provide an opportunity for direct intervention (eg, coronary revascularisation) that will likely have influence on the follow-up, treatment and prognosis of these patients.15 20 In the present study, performing coronary angiography was associated with improved outcomes, a finding that is consistent with the OPTIMIZE-HF registry11; however no causality can be established as this may reflect selection bias and better baseline patient profile. In a patient population with coronary artery disease and HF with reduced ejection fraction performing coronary artery bypass grafting (CABG) led to improved outcomes compared with medical therapy alone.21 Whether performing more coronary angiographies in patients with decompensated HF leads to outcome improvement needs to be prospectively evaluated in an adequately powered trial.
Older patients and those with worse renal function were less likely to have coronary angiography performed. It has been thoroughly documented that elderly patients and those with impaired renal function presenting with an ACS and/or acute HF undergo substantially less angiographic/revascularisation procedures, despite deriving similar relative benefits of these interventions.11 22 23 Remarkably, coronary angiography was not less likely to be performed in women, even though women in this study were older. Patients with cardiac devices, previous coronary interventions and HF hospitalisation, and those observed as outpatients were less likely to undergo coronary angiography. These findings may be due to the assumption that the patients were already investigated for coronary disease at the timing of device implantation or that those presenting as outpatients may have less severe symptomatology and require less investigation. Nevertheless, these patients may be at higher risk for myocardial ischaemia and stent restenosis.24
We found an ‘interaction’ between HF aetiology (ischaemic vs other) and the prognostic value of coronary angiography. Performing a coronary angiography in patients with non-ischaemic HF was associated with a better outcome than in patients with ischaemic HF. This finding is possibly due to the differences found in critical stenosis rates, which were much higher in patients with ischaemic HF (>80%) and were associated with worse prognosis. Patients who underwent coronary angiography and had coronary stenosis documented (≈39% in the present cohort) had worse prognosis compared with those without coronary stenosis. The presence of significant coronary lesions is associated with worse prognosis, as also documented in previous reports.25
Clinical and research implications
The present results show that coronary angiography was performed in <13% of patients with worsening HF. These subjects were younger and with a more favourable overall clinical profile. Therefore, these data should be taken as merely descriptive and no causality should be inferred from these observations. From a research standpoint, a trial comparing ‘usual care’ versus an arm with a low threshold for coronary angiography could provide more definitive answers on the diagnostic and prognostic abilities of this intervention.
Several limitations should be noticed in this study. First, this is a post hoc analysis of a prospective non-randomised observational study, therefore all limitations inherent to such analysis are applied herein, including the inability to infer causality. Additionally, it is likely that unmeasured variables may have accounted for the different outcomes observed. Second, the present study focused on changes in medication and determinants of medication uptitration (and not to address coronary angiography performance). Therefore, we unfortunately have no reliable and consistent information regarding the clinical consequences of the findings during coronary angiography. However, these data may reflect ‘real-world’ practices as no guidance was provided with regard to coronary interventions. Third, it is also impossible to account for the effect of indication biases that may have determined who underwent angiography as well as treatment biases that may have influenced who received pharmacological therapies for coronary artery disease and HF. Fourth, results from stress testing and/or coronary intervention outcomes (eg, stent placement, CABG referral) are not available in the data set. Fifth, the participating hospitals in the BIOSTAT-CHF differed in structure (from tertiary university hospitals to small non-academic structures) and likely in the access to a catheterisation laboratory, hence these findings cannot be generalised to all hospitals and patients with HF. However, further adjustment for the type of centre did not change the strength of the associations. Sixth, we can only hypothesise on the reasons that led clinicians to perform a coronary angiogram since this information is also not available. Lastly, the data from the BIOSTAT-CHF come from European centres only and may not be representative of patients with HF in other world regions.
Coronary angiography was performed in <13% of patients with symptoms and/or signs of worsening HF, particularly those presenting as inpatients, with an ACS, with better renal function and younger age. Performing a coronary angiogram was associated with improved outcomes but this observation possibly reflects a selection bias. These observations help inform the debate regarding the utility of coronary angiography for the investigation of worsening HF and might stimulate further research specifically to address this question.
What is already known on this subject?
Coronary angiography is regularly performed in patients with worsening signs and/or symptoms of heart failure (HF); however, previous reports show that less than 10% of these patients undergo coronary angiography.
What might this study add?
In our study, 12.5% of patients underwent coronary angiography within the 30 days after the onset of worsening symptoms and/or signs of HF. Subjects who underwent angiography were more likely to be inpatients, more likely to have an overt acute coronary syndrome, had higher troponin I levels, were younger and had better renal function. Patients who underwent coronary angiography had a lower risk of death and/or HF hospitalisation. Those with a coronary stenosis had a worse prognosis than those without stenosis.
How might this impact on clinical practice?
These observations help inform the debate regarding the utility of coronary angiography for the investigation of worsening HF and might stimulate further research specifically to address this question.
Contributors All authors contributed to the content of the present manuscript and approved its submission. JPF, AV and FZ designed the study, wrote the manuscript and provided critical appraisal. JPF performed statitical analysis. PR, BD, AS, NG, SDA, JGC, KD, GF, HLH, CCL, MM, LLN, PP, NJS, DJV and AHZ revised the manuscript.
Funding This project was funded by a grant from the European Commission (FP7-242209-BIOSTAT-CHF; EudraCT 2010-020808-29).
Competing interests None declared.
Ethics approval Ethics approval was obtained by each participating institution.
Provenance and peer review Not commissioned; externally peer reviewed.