Real-world Management of Heart Failure in the Philippines: The HF ER Registry

It is estimated that more than 64 million people worldwide have heart failure (HF), with most published studies reporting a prevalence of between 0.4% and 6.8% in the adult population.^1,2 In the Philippines, the prevalence approximates that reported in the US and Europe. Data from the Philippine National Insurance Corporation showed that the 2014 prevalence rate of hospitalised patients due to HF was 1.68% (1,648 cases of HF for every 100,000 patient claims for medical conditions).³

There had been several past registries of HF in the Philippines, but data on follow-up outpatient care after discharge were sparse in these registries.^4–7 Furthermore, data on rehospitalisation rates and long-term outcomes were also absent. However, these pieces of information are important in determining the real burden of disease and proposing improvements to healthcare delivery. Hence, there is a need to implement a high-quality HF registry that can produce real-world data that can improve patient care.

There are many reasons for the data gaps observed in past HF registries in the Philippines. These include (but are not limited to) patients transferring to other healthcare providers, which complicates follow-up; understaffing in institutions, which leads to the lack of manpower for data collection; poor understanding of data privacy laws; and other logistical issues related to data collection. One way to address these issues is through innovations and technologies, such as a well-funded purely electronic registry.

This observational prospective study aimed to determine the 12-month outcomes of HF patients enrolled in a purely electronic HF registry. The current analysis also aimed to describe the demographics, clinical characteristics and treatment patterns of HF patients in the Philippines.

Materials and Methods

Study Design, Setting and Included Patients

This HF registry is a pilot, non-interventional, longitudinal, multicentre registry of adult HF patients (≥18 years old, any sex, fulfilling the HF diagnosis based on the 2016 European Society of Cardiology guidance).² The protocol was reviewed and approved by the institutional review boards of the three participating institutions.

From October 2021 to September 2022, the study included patients 18 years and above diagnosed with HF based on the 2016 European Society of Cardiology guidelines on HF and seen at the Philippine General Hospital, Manila Doctors Hospital and Medical Center Manila—all tertiary medical centres. Patients should have typical signs and symptoms of HF and managed either as outpatients for chronic HF or admitted in the hospital due to decompensated HF. The investigators secured a written informed or digital consent before patient enrolment. Patients with HF with reduced, mildly reduced or preserved ejection fraction (EF) (HFrEF, HFmrEF and HFpEF, respectively) as well as uncategorised were all included.

Outcomes

After enrolment, all patients were instructed to follow up every 3 months for 1 year. However, data were available for most patients up for 400 days. The following information was obtained and recorded: symptoms and New York Heart Association classification; vital signs; echocardiographic findings including EF; medical history, including comorbidities; any laboratories done; medications, including dose; adherence to diagnostics and medications; any treatment adverse effect; any clinically important event; and enrolment to any clinical trial. The coprimary outcomes were acute decompensation, hospitalisation for HF, and death. Decompensation was defined as a worsening of symptoms requiring treatment with diuretics (or an increase in diuretic dose).

Data Management

The HF registry used an electronic medical record (EMR) application called BluEHR (Tantum Quantum Headquarters) for data collection. In compliance with the Data Privacy Act of 2012, only relevant non-identifying information was entered into the electronic registry. The EMR data were anonymised, hosted separately from the user application onto a cloud server, and protected by encrypted access. The EMR business logic is totally disconnected from the data to enforce anonymity and prevent data breaches. The EMR data are accessible only through the applications and administrative control system. No other access is available. Data entry is performed on a browser-based data entry website by the patient’s attending physician.

Statistical Analyses

The electronic registry provides easy data manipulation (with audit trails for reporting, monitoring and querying patient records). Patient demographics, clinical and radiological characteristics and medication tally were reported using descriptive statistics. Frequency and proportion were used for categorical variables. Shapiro–Wilk test was used to determine the normality distribution of continuous variables. Continuous quantitative data that met the normality assumption were summarised using mean and SD, while those that did not were described using median and range.

Continuous variables which were normally distributed were compared using the independent t-test (two-group comparison) or the one-way ANOVA (between three or more groups). Continuous variables which were not normally distributed were compared using the Mann–Whitney U-test (two-group comparison), or the Kruskal–Wallis H-test (between three or more groups). For categorical variables, χ² test was used to compare the outcomes. If the expected percentages in the cells were <5%, Fisher’s exact test was used instead. Kaplan–Meier survival analysis was used to analyse the clinical outcomes of acute decompensation, hospitalisation for HF, and death.

All valid data were included in the analysis. Missing variables were neither replaced nor estimated. Null hypothesis was rejected at 0.05 α-level of significance. Stata 15 was used for data analysis. The minimum sample size was 400 patients.

Results

The three study sites were able to enrol 394 HF patients into the registry. Table 1 shows the baseline characteristics of the included patients upon enrolment (on the first encounter, either at the emergency room or at the outpatient clinic). The majority were male (72.34%); the mean age was 56.91 ± 14.15 years. Only 75 patients (18.84%) had natriuretic peptide measurement. Regarding HF type, the majority had HFrEF (41.12%) or HFpEF (24.11%); however, 12.18% were unclassified. Coronary artery disease was the most common aetiology of HF (53.02%), followed by hypertensive heart disease (23.81%) and dilated cardiomyopathy (15.24%).

Table 2 shows that most patients were receiving a renin-angiotensin receptor (RAS) inhibitor (80.71%), β-blockers (71.57%) and two of the four drug classes of HF pharmacotherapy. However, a smaller proportion of patients were receiving the other two drug classes, namely mineralocorticoid receptor antagonists (MRA) (34.52%) and sodium-glucose cotransporter 2 inhibitors (SGLT2I) (28.93%). Overall, only 12.18% received four drug classes and 24.27% were receiving three drug classes. Supplementary Table 1 shows the detailed baseline medications.

Most patients included in the registry (n=334; 84.77%) had at least one follow-up visit. During the follow-up period, 61 patients died (400-day survival of 84.52%). The Kaplan–Meier survival curve for mortality is shown in Supplementary Figure 1a. There was a trend towards improvement of survival as the number of drug classes increased (Figure 1). Subanalysis showed that those who received angiotensin receptor/neprilysin inhibitor (ARNI) had a higher survival rate compared to those without ARNI treatment (p=0.030) (Supplementary Figure 1b).

During follow-up, 82 patients (20.81%) were hospitalised for HF (Supplementary Figure 2A). There was a trend towards a reduced risk of hospitalisation as the number of drug classes increased (Figure 2). The association between ARNI treatment and the risk of hospitalisation was not significant (0.091; Supplementary Figure 2B).

During follow-up, there were 26 decompensation events recorded, for a 400-day survival estimate of 0.9340 (95% CI [0.8797–0.9520]) (Supplementary Figure 3A). The risk of decompensation was significantly associated with a decreasing number of drug classes administered on enrolment (p=0.030; Supplementary Figure 3B). ARNI treatment showed a trend towards a lower risk of decompensation (p=0.170; Supplementary Figure 3C).

Lastly, the proportion of patients receiving at least three drug classes slightly increased from 33.74% at baseline to 50.94% at their last follow-up. The proportion of patients receiving four classes also increased from 11.81% at baseline to 20.05% (Supplementary Table 2).

Discussion

This pilot, non-interventional, longitudinal, multicentre registry of adult HF patients demonstrated that a purely electronic registry is implementable and may facilitate data collection for long-term registry data. In our cohort, the majority of HF patients had HFrEF (41.12%) or HFpEF (24.11%) and coronary artery disease was the most common aetiology of HF (53.02%). Data from the Global Burden of Disease study of 2017 showed similar trends, wherein the most common causes of HF globally were ischaemic heart disease (26.5%) and hypertensive heart disease (26.2%). Interestingly, the Global Burden of Disease study reported chronic obstructive pulmonary disease as the third most common cause of HF, contributing 23.4%.⁸ While the reason for this discrepancy is unclear, one possible reason could be the non-enrolment of such patients to the registry as these patients are likely being treated primarily for chronic obstructive pulmonary disease and rarely referred to cardiology care.

It should also be noted that 12.18% of patients enrolled in this real-world registry were unclassified due to lack of echocardiography results. This represents an important gap in the appropriate management of HF, as current guidelines recommend pharmacotherapy according to the level of EF. The absence of echocardiographic findings in over 12% of patients indicates that these patients were being managed suboptimally. The gap in the access and usage of echocardiography in the Philippines is well documented. The PHIL-SCREEN study published in 2021 showed that the Philippines faces challenges in echocardiographic resources and expertise nationwide. The study identified poor availability of echocardiography, inadequacy of expertise in echocardiography and other manpower issues as the main hindrances to the optimal use of echocardiography in the country.⁹ The authors of the present study also noted the cost of the test and long waiting lines as barriers to echocardiography in the country. The authors hope that this paper will encourage cardiologists and important stakeholders such as the Philippine Heart Association and the Philippine Society of Echocardiography to further advocate for routine echocardiography in all patients diagnosed with HF.

The authors also noted that only 18.84% of patients underwent natriuretic peptide measurement in real-world clinical practice – another important gap in the assessment of Filipino patients with HF. While all three study sites offer natriuretic peptide measurement, this test is not reimbursed under the national health insurance and is paid out-of-pocket. In government hospitals, this test costs around PHP1,200 (US$21), which represents twice the daily minimum wage;¹⁰ expectedly, the cost is higher in private institutions. These economic factors may have hindered the widespread adoption of natriuretic peptide measurement in the routine management of HF in the country. The authors of this paper call for the more action to increase the use of natriuretic peptide testing in patients with HF, as recommended by guidelines,¹¹ through physician education and improved access through reimbursement of government and private health insurers.

Additionally, other possible gaps in management include poor follow-up and low patient education, which may lead to poor treatment optimisation and low treatment adherence.

Guidance from the American Heart Association/American College of Cardiology recommends the use of four pillars of treatment (i.e. RAS inhibitors [ARNI, ACEI or angiotensin receptor blocker (ARBs), in order of preference], SGLT2I, MRAs and β-blockers) in the treatment of HFrEF, all class 1a recommendations.¹¹ Patients with HFpEF should also receive SGLT2I (class 2a recommendation), and MRAs and RAS inhibitors (ARNI or ARBs as an alternative) may also be considered as class 2b recommendations.¹² In our registry, the majority (80.71%) of patients in the registry were receiving an RAS inhibitor and 71.57% were receiving a β-blocker. However, ≤30% of patients were receiving an MRA or an SGLT2I. Overall, most enrolled patients (63.55%) were receiving less than three drug classes. Importantly, but not surprisingly, our analysis showed that the number of drug classes administered was associated with a lower risk of decompensation (p=0.030) or the composite of hospitalisation or death (p=0.001).

The reasons for non-adherence to any of the four drug classes were also recorded in the registry and will be discussed in detail in future work. Nonetheless, these findings indicate a major gap in the pharmacological management of Filipino patients with HF that needs to be addressed. Given the demonstrated benefit of these treatments on the outcomes of patients with HF, it is imperative to examine the reasons for non-adherence to guideline-directed medical therapy.

To our knowledge, this is the first HF registry conducted in the Philippines with robust follow-up data, with almost 85% of patients having at least one follow-up visit. We found that among Filipino patients with HF, 6.60% experienced acute decompensation, 20.81% were hospitalised, and 15.48% died during the follow-up period. These findings do not deviate far from global data. According to the Global Congestive Heart Failure study, which included data from 40 countries including several developing countries, while only 17% of patients experienced HF hospitalisation during the follow-up period, mortality was high at 19%.¹³ The EVOLUTION HF study, an observational, longitudinal cohort study using data from EMR or claims data sources from high-income countries, namely Japan, Sweden, the UK, and the US (2020–2022), found that 28% of patients died within the first year of HF hospitalisation.¹⁴ Patients in this study were enrolled at HF hospitalisation, indicating more severe disease and a contributor to the high mortality rate. However, the study strikingly found among those discharged, only 1.5% (US) to 11.3% (Sweden) were receiving all four HF drug classes. Time trends on the prescription rate of therapies showed that while the use of SGLT2I increased, the use of RAS inhibitors, β-blockers and MRAs remained stagnant. As the EVOLUTION HF study was conducted in developed countries and most have reimbursement for treatment, reasons other than healthcare resources are likely to play a role in the low usage of guideline-directed medical therapy.

The authors also noted that the results of the current study differ from major Asian HF registries. The China HF Registry study reported that echocardiographic LV function and natriuretic peptide test were conducted in 93.7% and 93.0% of the cases, respectively.¹⁵ These contrast sharply with our results showing around 12% have no echocardiographic testing and only 18% being tested for natriuretic peptide. The China HF registry also reported that 78.7% received an RAS inhibitor, 82.2% received a β-blocker, and 87.8% were on spironolactone. In contrast, our rates were 80.71%, 71.57%, and 34.52%, respectively.

The Trivandrum HF Registry from India reported contrasting results to our study. Their respective drug usage rates by drug class, depending on HF subtype, ranged from 43% to 50% (RAS inhibitors), 57% to 66% (β-blockers), and 31% to 51% (MRA).¹⁶ The study reported an unadjusted 5-year mortality rate of 59% and a crude 1-year survival rate of 66.39%. It is unclear why the reported 1-year survival in our study (84.52%) is higher, but this difference could be due to various reasons, such as differences in patient characteristics and study setting. It should be noted that the latest Trivandrum HF Registry paper reported results on patients enrolled from January 2013 to December 2013. In contrast, our study included patients enrolled from October 2021 to September 2022. During the interim, there were some developments in the treatment of HF. For example, EMPEROR-Reduced was presented in August 2020 and EMPEROR-Preserved was presented the following year.^17,18 Our study was then able to report that 28.93% of patients were already on SGLT2I. While this rate is far from ideal, the uptake of clinicians on the use of emerging pharmacotherapy in the management of HF is encouraging.

Description of the reasons for non-adherence to treatment is just one of the future analyses that should be undertaken on the robust data set available from the HF registry. Other planned analyses include multivariate analyses on prognostic factors, survival analyses on the various clinical outcomes, subgroup analyses on the adequacy of therapy and their impact on outcomes according to the HF type (HFrEF, HFmrEF and HFpEF), and detailed healthcare usage across the various subgroups. Furthermore, given that one limitation of the study is its implementation in only three institutions and with this pilot study being able to demonstrate the scalability of a purely electronic HF registry, more study sites may be included in the future.

Study Limitations

This is a non-interventional, non-randomised study; hence, adherence to treatment guidelines is affected by multiple patient- and physician-related real-world factors. While almost 84% of patients had at least one follow-up visit, the majority of patients did not follow up regularly with their physician, which may have affected not only the study’s data collection but also the quality of care they received. While only 12% of patients did not undergo echocardiography throughout the study, around 40% did not have echocardiographic results upon enrolment. This delay in performing an echocardiogram may also have impaired optimal treatment decision-making. Finally, the study only had a 1-year follow-up, which shows an incomplete picture of late mortality. Despite these limitations, the findings of this registry are an eye-opener for clinicians treating HF as the assessment and treatment of HF in the Philippines remains suboptimal and should be improved to ensure that patient outcomes are maximised.

Conclusion

This pilot HF registry showed that a purely electronic HF registry is implementable across various types of healthcare institutions and may be used to collect long-term data with a high rate of follow-up. Based on its preliminary results, most HF patients in the Philippines receive only suboptimal pharmacotherapy, which may contribute to the 21% hospitalisation rate and almost 16% mortality rate over a 400-day follow-up period.

Click here to view Supplementary Material.