Author + information
- Received August 10, 2018
- Revision received October 10, 2018
- Accepted October 14, 2018
- Published online December 31, 2018.
- Hans-Peter Brunner-La Rocca, MDa,∗ (, )@MaastrichtU,
- Gerard C. Linssen, MD, PhDb,
- Frank J. Smeele, MD, PhDc,
- Annemarie A. van Drimmelen, RNd,
- Henk-Jan Schaafsma, MDe,
- Paul H. Westendorp, MDf,
- Philip C. Rademaker, MDg,
- Hendrik J. van de Kamp, MSch,
- Arno W. Hoes, MD, PhDi,j,
- Jasper J. Brugts, MD, PhD, MSck,
- for the CHECK-HF Investigators
- aDepartment of Cardiology, Maastricht University Medical Center, Maastricht, Limburg, the Netherlands
- bDepartment of Cardiology, Hospital Group Twente, Almelo and Hengelo, Hengelo, the Netherlands
- cDepartment of Cardiology, Slingeland Hospital, Doetinchem, the Netherlands
- dDepartment of Cardiology, Amphia Hospital, Breda, the Netherlands
- eDepartment of Cardiology, Hospital Gelderse Vallei, Ede, the Netherlands
- fDepartment of Cardiology, Rivas Beatrix Hospital, Gorinchem, the Netherlands
- gDepartment of Cardiology, ZorgSaam Hospital, Terneuzen, the Netherlands
- hBusiness Unit Cardiometabolic, Servier Pharma, Leiden, the Netherlands
- iJulius Center for Health Sciences and Primary Care, Utrecht, the Netherlands
- jDepartment of Epidemiology, University Medical Center Utrecht and Utrecht University, Utrecht, the Netherlands
- kDepartment of Cardiology, Erasmus University Medical Center, Thoraxcenter, Rotterdam, the Netherlands
- ↵∗Address for correspondence:
Dr. Hans-Peter Brunner-La Rocca, Department of Cardiology, Maastricht University Medical Center, PO Box 5800, 6202 AZ Maastricht, the Netherlands.
Objectives This study investigated adherence to drug therapy guidelines in heart failure (HF) with reduced left-ventricular ejection fraction (LVEF) of <40% (heart failure with reduced ejection fraction [HFrEF]), in which evidence-based treatment has been established.
Background Despite previous surveys of HF, important uncertainties remain regarding guideline adherence in a representative real-world population.
Methods A cross-sectional registry in 34 Dutch HF outpatient clinics that included 10,910 patients with the diagnosis of HF was examined. Of that number, 8,360 patients had LVEF <50% (72 ± 12 years of age; 64% male) and were divided into HFrEF (n = 5,701), HF with mid-range LVEF (HFmrEF) with LVEF 40% to 49% (n = 1,574), and those with semiquantitatively measured LVEF but <50% (n = 1,085).
Results In the HFrEF group, 81% of the patients were treated with loop diuretics, 84% with renin-angiotensin-system (RAS) inhibitors, 86% with β-blockers, 56% with mineralocorticoid-receptor antagonists (MRA), and 5% with If-channel inhibition. Differences in medication use were minor among the 3 groups but were significant among centers. Inability to tolerate the medications was recorded in 9.4% patients taking RAS inhibitors, 3.3% taking β-blockers, and 5.4% taking MRAs. Median loop diuretic dose was 40 mg of furosemide equivalent, RAS inhibitor dose 50% of target, β-blocker dose 25% of target, and MRA dose 12.5 mg of spironolactone equivalent. Elderly patients were treated predominantly with diuretics and less often with RAS inhibitors, β-blockers, and MRAs.
Conclusions This large contemporary HF registry showed a relatively high use of evidence-based treatment, particularly in younger patients. However, the average dose of evidence-based medication was still lower than recommended by guidelines. Furthermore, the more recently introduced If-channel inhibition has hardly been adopted. There is ample room for improvement of HFrEF therapy, even more than 25 years after convincing evidence that HFrEF treatment leads to better outcome.
Adherence to guideline-recommended therapy in heart failure (HF) remains challenging. Adoption of new treatment options requires many years, often decades. Thus, angiotensin-converting enzyme (ACE) inhibitors were shown in the early 1990s to be beneficial in patients with heart failure with reduced ejection fraction (HFrEF) (1,2). However, more than 10 years later, only 60% of those patients were treated with an ACE inhibitor, although treating physicians were aware of the drug's benefits (3). The adoption of β-blockers, introduced later, was significantly worse (i.e., <40%) (4), and even currently mineralocorticoid receptor antagonists (MRAs) are still underused (5,6). At the same time, physicians' adherence to treatment guidelines is a strong predictor of favorable outcome (7,8). There is large regional variation both in Europe and in the United States, and this pertains not only to the use of HF medication but also device therapy (9).
Recent surveys found a higher prescription rate of evidence-based HF medication (8,10–12). Higher prescription rates were also reported in large treatment trials (13), but such patients are hardly representative of the general HF population, and inclusion requires evidence-based treatment. These patients are approximately 10 years younger than those in clinical practice, which also applies to some of the registries (8,12). Treatment uptake is lower in elderly patients, but guidelines recommend the use of HF medication regardless of age, despite the lack of clear evidence (14).
As shown in the recent U.S. CHAMP-HF (Change the Management of Patients With Heart Failure) registry (6) and in recent administrative data, the use of HF medication seems to be even lower in the real world than in recent surveys (8,10–12). Thus, patients included in these surveys were selected and were hardly representative of the general HF population. Moreover, registries often do not distinguish between patients with HFrEF and those with heart failure with preserved ejection fraction (HFpEF), although treatment recommendations to improve prognosis only apply to HFrEF patients (14). Also, many registries do not report the doses administered to patients (15). Although they are not well investigated, guidelines recommend using high doses of both renin-angiotensin-system (RAS) blockade and β blockade (14).
Because of these uncertainties, the present authors set up a cross-sectional registry of all patients seen in the outpatient clinic of 34 Dutch hospitals with the aim of investigating the quality of current HF management in the Netherlands. Because all patients treated by specialists in the Netherlands are seen in outpatient clinics of hospitals and not in private practices, this HF population was representative of patients seen by HF specialists.
Patients and clinics
The methods of the CHECK-HF (Chronisch Hartfalen ESC-richtlijn Cardiologische praktijk Kwaliteitsproject HartFalen) registry have been published previously (16). Briefly, HF or cardiology (if no specific HF clinic was present) outpatient clinics were invited to participate. Over a period of 3 years (2013 to 2016), 34 clinics (40%) of 86 centers in the Netherlands participated, of which 60 have an outpatient HF unit. Patients were included cross-sectionally based on the available records of these patients.
Diagnosis of HF was based on the most recent European guidelines for HF management available at the time (i.e., 2012) (17), but no further prerequisites were made regarding diagnostic and therapeutic decisions. Information included patients’ characteristics, main cause of HF, basic echocardiographic and electrocardiographic measurements, medication, co-morbidities, and relevant laboratory results. Patients' inability to tolerate and contraindication to the medication were recorded, as indicated by the treating physician. No predefined rules were applied to determine absolute contraindications. Ethical approval was provided for anonymously analyzing existing patient data by the Ethical Committee of the Maastricht University Medical Center, the Netherlands.
As part of the database, patients were divided based on left-ventricular ejection fraction (LVEF) into those with HFpEF with LVEF ≥50% and those with HF with LVEF <50%. In some patients, LVEF was only known semiquantitatively (using visual estimation) and divided into HFpEF versus HFrEF. In 283 (2.6%) of 10,910 patients, recording in the database was insufficient to classify patients according to their LVEF, and they were excluded from this analysis. Patients with no LVEF <50% were considered to have HFpEF (n = 2,267; 21.3% of the remaining patients). They were not considered for this analysis. Patients with LVEF <50% were divided based on the most recent European guidelines (18) into HFrEF (i.e., LVEF <40%), HF with mid-range LVEF (HFmrEF, i.e., LVEF 40% to 49%), and those with only semiquantitatively measured LVEF but whose LVEF was <50%.
Results are frequencies (%), mean ± SD, or median (interquartile range [IQR]), as appropriate. Between-group comparisons were performed using one-way analysis of variance (ANOVA), Kruskal-Wallis H test, or Pearson chi-square test, as appropriate. For multiple comparisons, the significance level was adjusted using a Bonferroni correction. Multivariate predictors of use of medication were sought, using multivariate logistic regression analysis, using the stepwise forward procedure. All predictors of medication use in univariate analysis (data not shown) at a p value of <0.10 were included in the multivariate regression analysis. Results of logistic expression are presented as odds ratio (OR) and level of significance.
A 2-sided p value of 0.05 was considered statistically significant. All calculations were performed using SPSS version 25.0 software (IBM, Armonk, New York).
Of the 10,910 patients included in the CHECK-HF registry, 8,360 patients (76.6%) with LVEF <50% were included in this analysis. Of these patients, 5,701 (68.2%) had HFrEF, 1,574 (18.8%) had HFmrEF, and 1,085 (13.0%) had only semiquantitatively estimated LVEF <50%.
Baseline characteristics of these patients are shown in Table 1. Patients were on average 72 years of age, and almost two-thirds were male. Most patients were in New York Heart Association (NYHA) functional class II, and only very few were in NYHA functional class IV. Coronary artery disease was the most common cause of HF, and most patients had relevant co-morbidities. HF had been diagnosed in most patients at least 1 year before inclusion. As shown in Table 1, there were some differences among patients with HFrEF, HFmrEF, and semiquantitatively measured LVEF. Most characteristics of the last group were somewhere between those of HFrEF and HFmrEF, suggesting a combination between the other 2 groups.
As shown in Figure 1, approximately 80% of the patients were treated with loop diuretics in all 3 groups (median dose of 40 mg furosemide equivalent; IQR: 20 to 80 mg). A similar proportion received a RAS inhibitor and a β-blocker. In HFrEF, 1.4% received a β-blocker not recommended by guidelines (1.0% received atenolol, 0.3% received propranolol, and 0.1% received labetolol 0.1%): 1.9% in the HFmrEF group and 1.6% in the semiquantitatively measured LVEF group, respectively. Sacubitril and valsartan were available only during the last couple of months of data collection and were given in only 0.4% of the patients (0.5% in HFrEF, 0.1% in the other 2 groups [analyses are included in RAS inhibitor results]). MRAs were prescribed to approximately one-half of the patients. Interestingly, an If channel inhibitor (ivabradine) was prescribed in <5% of the cases and was also low in those with LVEF <40% and with sinus rhythm (7.4%). Treatment significantly differed among the 3 groups for most medications. However, the absolute differences were relatively small. The combination of a RAS inhibitor and a β-blocker was given in 68.6%, 60.6%, and 63.1% in patients with HFrEF, HFmrEF, and those with semiquantitatively measured LVEF, respectively (p < 0.001). The combination of RAS blockade, β-blockade, and MRA was given in 39.2%, 27.3%, and 29.0% of patients, respectively (p < 0.001).
Patients who had had HFrEF diagnosed at least 1 year prior to inclusion, compared to those who had received their diagnosis within the same year, received loop diuretics less often (80.3% vs. 86.2%, respectively; p < 0.001), thiazides more often (2.9% vs. 1.4%, respectively; p = 0.03), equal percentages of β-blockers (80.7% vs. 83.6%, respectively; p = 0.09) and RAS inhibitors (82.6% vs. 83.2%, respectively; p = 0.63), MRAs less often (55.1% vs. 65.6%, respectively; p < 0.001), and equal percentages of ivabradine (5.2% vs. 6.8%, respectively; p = 0.10). Other drugs did not differ between the 2 groups. There was no specific pattern regarding these differences among the 3 groups (data not shown).
ACE inhibitors and angiotensin receptor blockers (ARB) were recorded as contraindicated in 9.4% of patients, with a wide range between centers from 0% to 36%. In 3.3% of the patients, β-blockers were recorded as contraindicated (range between centers of 0% to 27%), and in 5.4% of patients MRAs were contraindicated (range between centers of 0% to 22%). Thus, if this percentage is added to the prescription rates, 92.6% of the patients had either an ACE inhibitor or an ARB or were unable to tolerate them. The corresponding numbers were 84.3% for β-blockers (additionally 5.3% received sotalol), and 61.8% for MRAs. Because there were no major differences among the 3 groups based on LVEF, only numbers in HFrEF patients are given.
Prescription rates in HFrEF patients differed significantly (all p < 0.001) among the different centers (Figure 2A). Thus, all patients received loop diuretics in one center but only 63% in another. The largest differences were seen for MRAs, in which the prescription rates ranged between 34% and almost 90%. Also, the range for triple therapy was large, from 16% to 76%. Results in HFmrEF patients were even more pronounced (Online Figure S1).
Age significantly influenced the prescription rate in HFrEF patients as shown in Figure 3. Loop diuretics were more often prescribed to more elderly patients, whereas the opposite was the case for RAS inhibitors, β-blockers, and MRAs. Taking all patients into consideration did not change these results (data not shown).
HF medication predictions
Multivariate predictors of the use of HF medication are shown in Table 2. Only patients with complete information were included in the analysis (n = 4,043). Patients not considered for multivariate analyses due to lack of some data were somewhat less symptomatic (NYHA functional class I was observed in 21% vs. 12%, respectively; NYHA functional class II in 52% vs. 61%, respectively; remainder were the same), less often no co-morbidity (10% vs. 22%, respectively), and more often had LVEF that was only semiquantitatively measured (18% vs. 9%, respectively) or HFmrEF (21% vs. 17%, respectively), but furthermore did not differ in a clinically meaningful way from those included. Age and body mass index had influences on the prescription of all investigated drugs. As expected, LVEF influenced the prescription rate significantly. Finally, co-morbidity had some impact on prescription rate. The presence of coronary artery disease as underlying cause of HF did not influence the prescription rate (Table 2).
Target dose (for definitions see Online Table 1) and use of RAS inhibitors, β-blockers, and MRAs in patients with HFrEF are shown in Figure 2B: on average, MRAs and RAS inhibitors were given in higher doses than β-blockers. Still, approximately one-half of the HFrEF patients taking the prescribed medication were receiving less than the target dose of MRAs and RAS inhibitors. Considering all 3 groups based on LVEF, statistically significant but still relatively small differences regarding doses were seen (Online Figure 2). In all 3 groups, more than two-third of the patients received loop diuretics of 40 mg furosemide equivalent or less. Very high doses (i.e., >160 mg furosemide equivalent) were given in <5% of the patients.
A survey of this large, contemporary registry of HFrEF patients in a representative number of centers in the Netherland showed a high use of evidence-based HF treatment. This was particularly true in younger patients but significantly less in more elderly patients who were treated predominantly with diuretics; yet, the average doses used for both RAS inhibitors and particularly β-blockers are still much lower than those recommended by the guidelines (14,18). MRAs were less used, and more recently introduced treatment such as If channel inhibition (19) was hardly adopted. There was a large variation in the use of evidence-based therapy among different centers; medications not recommended by guidelines were used (e.g., sotalol, atenolol); and LVEF was only semiquantitatively estimated in several patients. Therefore, and in line with the recent U.S. CHAMP-HF registry (6), there is still room for improvement of HFrEF therapy, despite the abundance of evidence for how to manage HFrEF patients.
A report by the Institute of Medicine suggested that it takes on average 17 years before new knowledge generated in randomized trials is incorporated into practice and even then acceptance varies considerably among centers (20). The difficulty in efficiently adopting new treatment has recently been confirmed by the Get-With-the-Guidelines Heart Failure registry for the introduction of sacubitril/valsartan (21) and even more impressively by the very recent CHAMP-HF registry (6). Better adoption of recommended therapy could theoretically result in a substantial reduction in HF deaths (22), as adherence to guideline-recommended medical therapy resulted in better outcome independent of device therapy (23). Also, a recent registry showed better short-term outcome with adherence to good general guidelines (8). Our data suggest that contemporary use of evidence-based treatment in HF is also reasonable in a real-world setting, but significant further improvement is required (6). Still, it is difficult to understand how such improvement can be achieved (20). Reasons for the limited adoption of not only new but also established therapy and the variation in practice are hardly understood. It is only relatively recently that the absolute number of recommended therapies as well as the possible reasons for not using such treatments have been investigated. However, usually such assessments are based on subjective opinion by the treating physician, as in our study. It is well recognized that reasons for not giving certain drugs may vary significantly among physicians and that drugs may be perceived as contraindicated even if they are not (24). Therefore, it is not surprising that not only did the prescription rates vary significantly but also the patients' recorded inability to tolerate the drugs. This highlights the need to also assess reasons for changing therapy in more detail, as was done in the CHAMP-HF registry (20). Various factors for not making optimal use of medication could be identified, and some of them represent accepted side effects (6). However, many factors were identified that were not considered reasonable for underuse. It would be important to record not only specific instances of the inability to tolerate a drug but also to independently evaluate the individual decisions for not following guidelines. As far as these authors are aware, such an approach has not yet been followed. Although this kind of investigation is very time consuming, it would help to understand the real motivation for not strictly following the guidelines. A better understanding of health care provider and patient barriers to adoption of evidence-based therapies into routine clinical practice would have tremendous public health implications by designing effective quality improvement interventions.
Another possibility to improve adherence to therapy recommendations is benchmarking between different institutions and/or caregivers. This may result in in-depth reflection of therapeutic decision making and in sharing shortcomings and possible solutions. This is one of the important aims of the CHECK-HF registry (16). Indeed, the present authors found interesting and quite large differences among the participating centers, which will be fed back to all centers individually. Because the start of a Dutch national registry for HF has been planned for the near future, the impact of such feedback to adopt best clinical practice can be tested in the future.
To some extent, prescription of medication that does not accord with the guidelines is related to patient age. This may be seen as justified, at least to some extent, as in large randomized clinical trials, patients older than 75 years of age were rare with few exceptions (25). Thus, the positive effects in the very elderly patients and those with significant co-morbidities have not yet been sufficiently investigated. Although medication prescription was significantly influenced by age in multivariate analysis, it was much less influenced by co-morbidities. Also, given the large variations among the centers, it is unlikely that consideration of patient age is the main reason for not following the current guidelines.
Interestingly, treatment of patients with HFmrEF did not differ much from that in patients with HFrEF. This is in line with a recent report from the European Society of Cardiology Heart Failure Long-Term Registry in a population that was younger than that in this registry (26). In a recent report of a biomarker-guided trial including an older population, differences in baseline treatment between HFrEF and HFmrEF patients were also in a similar range (27). Differences were smaller than the variations in treatment seen among participating centers in this study. Therefore, indications for treatment in HFmrEF seem to be little different than that in HFrEF, although there is little support from prospective studies. Still, very recently, there is some evidence from a post-hoc analysis of the CHARM (Candesartan in Heart failure Assessment of Reduction in Mortality and morbidity) trial that ARB has an effect in HFmrEF that is similar to that HFrEF (28). A subgroup analysis of the TOPCAT (Treatment of Preserved Cardiac Function Heart Failure with an Aldosterone Antagonist) study suggested that MRAs are beneficial in patients with slightly reduced LVEF (i.e., in the range of HFmrEF) but less or not in patients with fully preserved LVEF (29). These 2 post hoc analyses provide some support that the assumption of similar treatment responses in HFrEF and HFmrEF may be correct. However, this needs to be tested prospectively. In this regard, it is important that LVEF is precisely determined and not only semiquantitatively as done in some patients in this registry, which may contribute to a large variability in the evaluation of LVEF (30). Also, other therapies such as implantable cardioverter-defibrillator insertion are dependent on exact measurements of LVEF (14,18).
No follow-up data are available; therefore, it is possible that some patients were still in the uptitration phase, and final use of recommended therapy could have been closer to the guidelines than reported in this study. However, time of diagnosis did not influence the use of medication significantly. In addition, only approximately one-half of the centers with an outpatient HF clinic and 1 of 8 university centers participated. Still, this percentage is much higher than in most previous registries. Therefore, the reported data may reasonably reflect the current situation in the Netherlands.
Moreover, data were reported at the discretion of the participating centers, recorded based on existing patient files, and some data are therefore missing. We cannot exclude the possibility that this might have influenced our results. However, differences among patient characteristics did not differ in a clinically meaningful way among patients where information was available and where it was not. Moreover, differences among centers cannot be explained by missing values. Furthermore, this registry included only patients seen in secondary but not in primary care. This may explain why some of the characteristics were not fully in line with unselected populations. Still, our population represents an older population than that in many other registries and is a large and representative sample of Dutch HF patients in secondary care. In addition, patient selection by the necessity to obtain informed consent is absent, which is an important strength of this registry, improving the reflection of real-world information. Interestingly, the results of this registry are in many ways in line with the very recent CHAMP-HF registry in the United States that also included a representative HF population (6). Still, the overall use of medication and the doses used in CHAMP-HF were slightly lower than those in CHECK-HF. In contrast to CHAMP-HF, hardly any information is available for the use of sacubitril/valsartan since it was approved in the Netherlands only in June 2016. Also, the use of isordil/hydralazine is so low in the Netherlands that the information was not collected. From the American and European perspectives, both of the large registries CHAMP-HF (6,20) and CHECK-HF indicate the continuing need for optimizing guideline adherence and the importance of understanding reasons for nonadherence.
In a contemporary cohort of HF patients with reduced and mid-range LVEF, most patients received guideline-recommended medication but at doses lower than recommended. Moreover, ivabradine, which was introduced more recently was prescribed less. Patient age influenced prescription rates of medication significantly. Importantly, there was a large variation among centers, which could not be explained by differences in patients’ characteristics. Medication in HFmrEF patients did not differ much from that in HFrEF patients, and LVEF was not precisely measured in a substantial number of patients. Therefore, there is still significant room for (further) improvement of treatment of patients with reduced LVEF, even 30 years after the first study showing that prognosis can be improved in HFrEF (31).
COMPETENCY IN MEDICAL KNOWLEDGE: Results of this registry indicate that much more effort should be placed to understand the incomplete adoption of guideline recommendations in clinical practice.
TRANSLATIONAL OUTLOOK: Only if the effective measures are understood can they be taken to improve outcome. Such measures may include new means to educate healthcare professionals and to effectively involve patients in the treatment process, but also research on a more personalized approach of treating HF patients to more specifically target treatment to the needs of individual patients.
The authors thank the HF nurses and cardiologists of all participating sites for their participation in including patients and recording patient data.
Data inclusion and software were supported by Servier. There was no funding source for analysis, data interpretation, and writing of the manuscript. Dr. Brunner-La Rocca has received research grants from Roche Diagnostics, Novartis, and Vifor. Dr. van de Kamp is an employee of Servier. All other authors have reported that they had no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- angiotensin-converting enzyme
- angiotensin receptor blocker
- heart failure
- heart failure with mid-range ejection fraction
- heart failure with preserved ejection fraction
- heart failure with reduced ejection fraction
- left-ventricular ejection fraction
- mineralocorticoid receptor antagonist
- renin-angiotensin system
- Received August 10, 2018.
- Revision received October 10, 2018.
- Accepted October 14, 2018.
- 2019 American College of Cardiology Foundation
- Pfeffer M.A.,
- Braunwald E.,
- Moye L.A.,
- et al.
- Savarese G.,
- Carrero J.J.,
- Pitt B.,
- et al.
- Greene S.J.,
- Butler J.,
- Albert N.M.,
- et al.
- Komajda M.,
- Cowie M.R.,
- Tavazzi L.,
- Ponikowski P.,
- Anker S.D.,
- Filippatos G.S.
- Shah B.,
- Hernandez A.F.,
- Liang L.,
- et al.
- Maggioni A.P.,
- Anker S.D.,
- Dahlstrom U.,
- et al.
- Crespo-Leiro M.G.,
- Anker S.D.,
- Maggioni A.P.,
- et al.
- Yancy C.W.,
- Jessup M.,
- Bozkurt B.,
- et al.
- Brugts J.J.,
- Linssen G.,
- Hoes A.W.,
- Brunner-La Rocca H.P.,
- for the investigators C-H
- McMurray J.J.,
- Adamopoulos S.,
- Anker S.D.,
- et al.
- Ponikowski P.,
- Voors A.A.,
- Anker S.D.,
- et al.
- DeVore A.D.,
- Thomas L.,
- Albert N.M.,
- et al.
- Luo N.,
- Fonarow G.C.,
- Lippmann S.J.,
- et al.
- Roth G.A.,
- Poole J.E.,
- Zaha R.,
- Zhou W.,
- Skinner J.,
- Morden N.E.
- Chioncel O.,
- Lainscak M.,
- Seferovic P.M.,
- et al.
- Rickenbacher P.,
- Kaufmann B.A.,
- Maeder M.T.,
- et al.
- Lund L.H.,
- Claggett B.,
- Liu J.,
- et al.
- Kaufmann B.A.,
- Min S.Y.,
- Goetschalckx K.,
- et al.