Clinical profile and 90 day outcomes of 10 851 heart failure patients across India: National Heart Failure Registry

Abstract Aims Limited data on the uptake of guideline‐directed medical therapies (GDMTs) and the mortality of acute decompensated HF (ADHF) patients are available from India. The National Heart Failure Registry (NHFR) aimed to assess clinical presentation, practice patterns, and the mortality of ADHF patients in India. Methods and results The NHFR is a facility‐based, multi‐centre clinical registry of consecutive ADHF patients with prospective follow‐up. Fifty three tertiary care hospitals in 21 states in India participated in the NHFR. All consecutive ADHF patients who satisfied the European Society of Cardiology criteria were enrolled in the registry. All‐cause mortality at 90 days was the main outcome measure. In total, 10 851 consecutive patients were recruited (mean age: 59.9 years, 31% women). Ischaemic heart disease was the predominant aetiology for HF (72%), followed by dilated cardiomyopathy (18%). Isolated right HF was noted in 62 (0.6%) participants. In eligible HF patients, 47.5% received GDMT. The 90 day mortality was 14.2% (14.9% and 13.9% in women and men, respectively) with a re‐admission rate of 8.4%. An inverse relationship between educational class based on years of education and 90 day mortality (high mortality in the lowest educational class) was observed in the study population. Patients with HF with reduced ejection fraction and HF with mildly reduced ejection fraction who did not receive GDMT experienced higher mortality (log‐rank P < 0.001) than those who received GDMT. Baseline educational class, body mass index, New York Heart Association functional class, ejection fraction, dependent oedema, serum creatinine, QRS > 120 ms, atrial fibrillation, mitral regurgitation, haemoglobin levels, serum sodium, and GDMT independently predicted 90 day mortality. Conclusion One of seven ADHF patients in the NHFR died during the first 90 days of follow‐up. One of two patients received GDMT. Adherence to GDMT improved survival in HF patients with reduced and mildly reduced ejection fractions. Our findings call for innovative quality improvement initiatives to improve the uptake of GDMT among HF patients in India.


Introduction
Heart failure, a complex clinical syndrome often arising from impairment of ventricular filling or ejection of blood, is estimated to affect 26 million people worldwide. 1 It is disabling and deadly, with an annual hospitalization of approximately one million US adults 2 and a 1 year mortality of 23.6% in patients hospitalized with acute HF in high-income settings. 3 Heart failure prevalence varies across regions of the world. 4,5 However, population-based data on HF prevalence and incidence from India are scarce. Data from the Trivandrum Heart Failure Registry, 6-8 the International Congestive Heart Failure registry, 9 and the Medanta registry 10 provide crucial information on patient characteristics, prevailing treatment practices, and survival of HF in India. However, all of these registries are limited in their geographical representations of India. 11 The National Heart Failure Registry (NHFR) tries to bridge the gap with representation from diverse geographical areas of India. We describe the demographic characteristics, disease aetiology, practice patterns, 90 day mortality, and predictors of 90 day mortality of acute decompensated HF (ADHF) patients from the NHFR.

Study settings
The design and detailed methods of the NHFR have been published earlier. 12 In brief, the NHFR is a facility-based registry of ADHF patients admitted to 53 hospitals in 21 states and four union territories in India. The overall recruitment period was from January 2019 to July 2020. Nine nodal centres that were government-funded medical institutes with adequate experience in conducting epidemiological studies were selected from nine different regions in India. Each nodal centre identified five additional hospitals, representing the different geographies and ethnicities in the region as participating sites. The 53 centres included in the NHFR ensured adequate representation of different geographical regions of India (see Supporting Information, Figure S1).

Participants
All consecutive ADHF patients >18 years who satisfied the European Society of Cardiology (ESC) 2016 criteria 13 were enrolled in the registry. All included patients had an echocardiogram during the hospitalization period.

Data collection and definitions
Data on socio-demographic characteristics, aetiology of HF, medical history, medications, clinical presentation, in-hospital diagnosis, and treatment were retrieved from medical records and interviews of the patients or their immediate family members. Trained staff nurses collected relevant data and uploaded them to the NHFR central server using either an Android-based or a web-based data entry application. Data uploaded to the NHFR server were verified for accuracy and completeness by the nodal centre co-ordinator. The data quality was monitored using centrally generated periodic listing of data queries. All related data queries were reviewed and resolved by the nodal centre co-ordinators with the help of the National Coordinating Centre at Sree Chitra Tirunal Institute for Medical Sciences and Technology (SCTIMST), Trivandrum, India. Further, 10% of data fields were subjected to source data verification, and all participating sites reported <1% data errors.

Longitudinal follow-up
The site co-ordinators or research nurses at each site followed up all patients enrolled in the NHFR for a minimum period of 3 months from the date of admission. The follow-up data were collected either during the clinic visits or by telephone. Data related to repeat hospitalizations, mortality, and cause of death were collected directly either from the patients or from their immediate family members. The date of such events was also collected and recorded as part of the study. In the case of out-of-hospital mortality, the research staff completed a verbal autopsy to ascertain the cause of death.

Outcomes
All-cause mortality at 90 days was the main outcome variable. We used the time-to-mortality data based on the date of event in the survival analysis. parametric hypothesis testing. Continuous variables are presented as means with standard deviation (SD). We used one-way ANOVA for comparison of means across the three HF categories.
We reported in-hospital, 90 day mortality rate as a proportion (number of deaths/total registered patients) and per 100 person-days of follow-up. Further, we employed survival analyses to identify factors associated with 90 day mortality outcomes. Kaplan-Meier survival models were performed initially, after checking the proportional hazards' assumption by using log-minus-log plots, and groups were compared using log-rank tests. Finally, a multi-variable Cox proportional hazard model (Cox-PH) was employed to assess the adjusted hazard ratio (HR) of all-cause mortality. We used variables that are associated with mortality in previous models of acute HF 17,18 and are available in the NHFR data set to generate the regression model. All analyses were carried out using Stata 16.1 StataCorp LLC.

Ethical considerations
We obtained institutional ethics committee approval from all participating centres before the commencement of the study. The Institute Ethics Commitee of the co-ordinating centre (SCTIMST) also cleared the study (no.: SCT/IEC/1167). We registered this study in the Clinical Trial Registry of India (CTRI/2019/01/017053). A written informed consent was obtained from all participants before enrolment into the registry.

Funding
We received a research grant from the Indian Council of Medical Research (ICMR), Department of Health Research, Government of India, for the conduct of this study.

Results
A total of 10 851 patients were recruited. The mean age of the study population was 59.9 (SD = 13.5) years ( Table 1). Women contributed to less than one-third (31%) of the study population. The mean year of schooling was 7.8 (SD = 5.2) years. One-fourth of the study population (25.5%) reported less than 4 years of formal schooling.
The mean body mass index (BMI) of the study population was 24.2 (SD = 4.0) kg/m 2 . More than three-fourths (74.4%) of the patients had admission with de novo HF, whereas the remaining (25.6%) patients had re-admissions. The most common presentation was HFrEF in two-thirds (65.2%) of the study population, followed by HFmrEF (22%) and HFpEF (12.7%). Ischaemic heart disease was the predominant aetiology for HF (72%), followed by dilated cardiomyopathy (18%). Rheumatic valvular heart disease (RHD) was present in 5.9% of the study population. Valve diseases other than RHD contributed 2.1% of the cases. Isolated right HF was noted in 62 (0.6%) participants.
Hypertension and diabetes were the most frequent co-morbid conditions (48.9% and 42.3%, respectively). Atrial arrhythmia was prevalent in 9.5% of the study population. Chronic kidney disease, stroke, and chronic obstructive pulmonary disease (COPD) were reported in 8.5%, 3.0%, and 6.9% of the study population, respectively. Similarly, tobacco and alcohol use were reported in 33.8% and 16.7% of the study population, respectively.
More than two-thirds (71.2%) of the patients presented with New York Heart Association (NYHA) Class III or Class IV. Further, about one-third (35%) reported a baseline heart rate of more than 100 beats/min. A baseline systolic blood pressure of <90 mmHg and a diastolic blood pressure of <60 mmHg were reported in 4.6% and 5.7% of the study population, respectively. The mean serum creatinine was 1.5 (SD = 1.1) mg/dL, whereas the mean haemoglobin was 12.1 (SD = 2.2) g/dL. Serum sodium < 125 mEq/L and serum potassium > 5.5 mEq/L were reported in 3.3% and 4.5% of the study population, respectively.

Prescription pattern and outcomes
During hospital admission, non-invasive and invasive ventilatory support were required in 17.6% and 3.9% of the study population, respectively. Intravenous inotropes were administered in one-fifth (20.2%) of the patient population. In HFrEF, 47.5% received GDMT ( Figure 1). Similarly, more than one-third of HFmrEF also received GDMT. The ARNI use was very low in HFrEF (4.8%) and HFmrEF (4.7%) patients.
Overall, 90 day mortality was 14.2% (14.9% and 13.9% in women and men, respectively). In-hospital mortality was 6.7% (Table S1). It was highest in HFrEF (7.5%) and lowest in HFmrEF (5.1%). In-hospital mortality was 7.6% and 6.7% in women and men, respectively. Patients with HFrEF and HFmrEF who did not receive GDMT experienced higher mortality (log-rank P < 0.001) than those who received it ( Figure 2). The 90 day mortality in HFmrEF was higher in women than in men (see Supporting Information,  Figure S2). The re-admission rate during the 90 day follow-up was 8.4% in the study population. An inverse relationship between educational class based on years of education and 90 day mortality (high mortality in the lowest educational class) was observed in the study population ( Figure 3A). The strength of the association did not attenuate in the multi-variable Cox proportional hazard model ( Figure 3B). Further, baseline BMI, EF, NYHA class, dependant oedema, mitral regurgitation, QRS duration > 120 ms, presence of COPD, atrial fibrillation, use of GDMT, dilated cardiomyopathy, haemoglobin, serum sodium < 125 mEq/L, use of intravenous inotropes and diuretics, and respiratory support during admission were associated with 90 day mortality ( Table 3).
Patients with NYHA Classes III and IV at the time of admission had 36% [HR = 1.36; 95% confidence interval Individuals with more than 13 years of schooling reported 37% (HR = 0.63; 95% CI: 0.52-0.75) lower 90 day mortality as compared with individuals with <4 years of schooling. The presence of dependent oedema and atrial fibrillation was associated with 21% and 29% higher mortality as compared with individuals without such clinical signs at the time of admission. Broad QRS in electrocardiogram was associated with 40% (HR = 1.40; 95% CI: 1.26-1.56) higher mortality. Each unit increase in haemoglobin in grams per decilitre was associated with 5% lower 90 day mortality (HR = 0.95; 95% CI: 0.93-0.97). Lower serum sodium below 125 mg/dL at the time of admission was associated with 61% higher mortality (HR = 1.61; 95% CI: 1.30-1.98) as compared with individuals with serum sodium above 125 mg/dL. The requirement of both non-invasive and invasive ventilation was associated with higher mortality as compared with individuals without the requirement of respiratory support.

Discussion
Our findings based on the NHFR, the largest HF registry from India with adequate representation of all regions, provide Figure 1 Intake of key drugs in men and women.ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker; ARNI, angiotensin receptor-neprilysin inhibitor; BB, beta-blocker; HFmrEF, heart failure with mildly reduced ejection fraction; HFpEF, heart failure with preserved ejection fraction; HFrEF, heart failure with reduced ejection fraction; MRA, mineralocorticoid receptor antagonists. CCB, calcium channel blocker; HF, heart failure; HFmrEF, heart failure with mildly reduced ejection fraction; HFpEF, heart failure with preserved ejection fraction, HFrEF, heart failure with reduced ejection fraction; LMWH, low-molecular-weight heparin; OAC, oral anticoagulants.
crucial information on patient characteristics, treatment practices, and 90 day mortality outcomes of HF patients. In our cohort of relatively young patients with a mean age of 60 years, ischaemic heart disease was the predominant disease aetiology in three of four ADHF patients. Approximately one of two patients with HFrEF received GDMT. One of seven patients died during the first 90 days of follow-up from the date of admission. Educational class, dependent oedema, QRS > 120 ms, atrial fibrillation, mitral regurgitation, haemoglobin levels, and serum sodium at baseline predicted 90 day mortality over and above the traditional risk score variables. The hospitals included in the NHFR were spread across different regions of India with adequate coverage to the northern, southern, western, and eastern parts of India. Further, it covers states in all regions based on the epidemiological transition levels. 19 The educational level of participants in the NHFR also reflects the national average educational level in this age group. As noted in several other registries, 6,9,10 HF patients in India are younger than HF patients from high-income countries. 3,20 Further, the mean age was a decade younger than the patients in the heart failure registry of patient outcomes 21 study from China.
Male preponderance in HF incident cases as noted in the NHFR is a common observation in patients from India. 6,9 However, patients from HF registries in high-income regions show a relatively high proportion of women. 3,20 High incidence of ischaemic heart disease in men may be one of the reasons for male preponderance as it is the commonest disease aetiology in the NHFR. However, a lack of tertiary care access for women in India has been reported earlier and may be responsible for the missing female HF patients. 22 Heart failure equally affects the low and high socio-economic strata in India. For example, one of four HF patients in the NHFR reported either no or less than 4 years of formal schooling. It is also striking that the 90 day HF mortality is highest in individuals with low educational status with a clear inverse relationship between educational level and mortality. A similar association has been noted in acute coronary syndrome registries 23 and several other studies in India. [24][25][26] Surprisingly, this inverse relationship persists in the NHFR even after adjustment for traditional factors associated with mortality including GDMT. A relatively high mortality in women, especially in the group who did not receive medications as per GDMT is also a cause of concern. The socio-economic differentials in HF mortality in India need to be further investigated to provide clear recommendations for management in this high-risk group of individuals.
In our registry, almost one of two patients with HFrEF received GDMT. This is a significant improvement based on the rates reported from previous HF registries in India. 6,10,27 The setting up of registries and dissemination of its results may itself improve adherence to GDMT. The prescription rate of GDMT in the NHFR is comparable to that in acute HF patients in the ESC HF Long-Term Registry, 3 which provides data from six different regions of Europe. The uptake of GDMT in various HF registries is summarized in Supporting Information, Table S2. It is also important to note that those who received the GDMT survived longer than those who did not receive it during the initial hospital stay. The benefits of GDMT accrued in HFrEF were also noted in HFmrEF. This is consistent with findings from other studies that show that the combinations of ACEI/ARB, betablocker, and aldosterone blocker benefit HFmrEF patients with a reduced mortality. 28 Consistent with previous findings from India, the use of ARNI and devices like CRT and ICD was very low in the NHFR. The device treatment rate in the NHFR is at least nine times lower than that reported in high-income countries, 29 but similar to recent data from Figure 2 Guideline-directed medical therapy, heart failure groups and 90 day survival.GDMT, guideline-directed medical therapy; HFmrEF, heart failure with mildly reduced ejection fraction; HFpEF, heart failure with preserved ejection fraction; HFrEF, heart failure with reduced ejection fraction. Tunisia. 30 Issues related to affordability and accessibility to such treatment could be a major barrier in improving the use of such expensive but lifesaving treatment options in the Indian context. 31 One of seven patients in the NHFR died during the first 90 days of follow-up with an in-hospital mortality of one in 15 patients. In-hospital mortality rate in the NHFR was almost two-fold higher than that in the 'ESC HF Long-Term Registry' of acute HF patients. 32 However, both in-hospital and 90 day mortality rates in NHFR were lower than that reported previously in HF registries from India. 6 Wider knowledge dissemination of benefits of the GDMT and its uptake may have improved the overall survival of HF patients in the NHFR. This is also consistent with the observation on relatively low mortality in HF patients who received GDMT in the NHFR. However, the mortality rate in registries from high-income regions is even lower than that reported in the NHFR. 33 It suggests that there is further scope for im- provement and calls for innovative methods to improve the uptake of GDMT in eligible patients in India to reduce the mortality associated with HF.

Strengths and limitations
National representation and data from all regions of India are the key strengths of our study. A uniform study protocol was followed at all sites for data collection with central monitoring for possible data errors. The lack of central adjudication committee to confirm the causes of death and type of re-hospitalization is a major limitation of our study, and therefore, we restricted our analysis to all-cause mortality. The registry included only patients from tertiary hospitals and departments headed by cardiologists, and the extent to which the findings from this study can be generalized to patients admitted in other hospitals with limited facilities is unclear. The new additions to GDMT such as the sodium-glucose co-transporter 2 inhibitors were not approved at the time of enrolment in the NHFR.

Conclusions
Ischaemic heart disease is the leading cause of HF in India in a younger cohort of patients from the NHFR. Almost one of two eligible patients received guideline-directed therapy in the NHFR. GDMT is associated with improved survival of HFrEF and HFmrEF. The 90 day mortality rate of one in seven patients and the sub-optimal uptake of GDMT call for nationwide quality improvement initiatives.

Conflict of interest
None declared.

Supporting information
Additional supporting information may be found online in the Supporting Information section at the end of the article. Table S1: Mortality outcomes. Table S2: Prescription of disease modifying agents in heart failure. Figure S1: Study sites in the NHFR. Figure S2: Heart failure sub-groups and 90-day mortality. Appendix S1. Supporting Information.