This study was conducted in Catalonia, an autonomous region in north-east Spain with 7.76 million inhabitants as of January 2022. Data were obtained from the Catalan Health Surveillance System (CHSS), a population-based registry managed by the Catalan Health Department.19,20 The CHSS compiles demographic, socioeconomic, diagnostic, and health care utilization data for all individuals covered by the public health care system. It integrates information from multiple sources, including the Minimum Dataset for Health care Units (hospitalizations, primary care, skilled nursing, outpatient specialist visits, and emergency care), pharmacy prescription records, and billing data. Until 2017, diagnoses were coded using ICD-9-CM, ICD-10-CM, or ICPC-2, depending on the care setting. From 2018 onwards, the ICD-10-CM was adopted in hospitals and primary care. All diagnoses were harmonized to ICD-10-CM, procedures were coded using ICD-9 until 2017 and ICD-10 thereafter. Medications were classified according to ATC codes, and vital status was obtained from the Spanish National Statistics Institute.

In this retrospective, population-based study, we analyzed data from the CHSS between January 2011 and December 2021, covering an 11-year period. The study was registered at ClinicalTrials.gov (NCT06255418) as BEAT-HF (Big data & real-world Evidence to assess the healthcare and health outcomes burden of Acute coronary syndromes complicaTed with Heart Failure). Ethics approval was obtained from the Clinical Research Ethics Committee of the Bellvitge Biomedical Research Institute, Spain, and the study complied with the Declaration of Helsinki. Owing to its retrospective design and use of pseudonymized data, informed consent was waived.

We included data from all residents of Catalonia covered by the public health care system who were diagnosed with ACS during a hospitalization episode due to NSTEMI, STEMI, or UA and were discharged between January 1, 2011 and December 31, 2021 according to the prespecified diagnostic coding described in table S1. We excluded patients who had been diagnosed with HF in the years preceding the study period and those who died before discharge. iHF was defined using the following criteria: a) first hospitalization with a primary diagnosis of HF after the index ACS (table S2), or b) first HF diagnosis recorded in primary care after the index ACS plus concurrent prescription of loop diuretics to validate the diagnosis.

The first coprimary objective was to describe the baseline characteristics of the surviving ACS population overall and according to the event type (STEMI, NSTEMI, UA) and other relevant subgroups of patients. The second coprimary objective was to investigate the real-world incidence of HF after ACS in all patients discharged after ACS. The secondary objectives of the study were: a) to ascertain the occurrence of important clinical outcomes after ACS, b) to describe the medical resource use and expenditure of the overall population of patients experiencing an ACS, and c) to analyze these data stratified according to iHF status.

We evaluated the following clinical outcomes: occurrence of all-cause mortality and need for CV procedures. Additional outcomes assessed comprised the occurrence of recurrent ACS, stroke, CV hospitalization, 4-point major adverse cardiac event (MACE) (all-cause mortality, recurrent ACS, new coronary revascularization, and stroke), and 5-point MACE (4-point MACE plus iHF).

We analyzed health care resource use (emergency, daycare, outpatient specialist, and primary care contacts, skilled nursing, and unplanned hospitalizations, including HF, CV, clinically related, and all-cause admissions) and associated costs (global, category-specific, and drug expenditure). Health care expenditure was calculated as previously described.21 Patients’ socioeconomic status was determined according to medicine copayment levels, as detailed elsewhere.19 Killip classification was used to assess ACS severity, ranging from no HF (Killip I) to cardiogenic shock (Killip IV). Clinical complexity was assessed using the Adjusted Morbidity Groups, which classify the population into 6 morbidity groups and have been validated in the Spanish and Catalan public health care systems.19–22

Baseline characteristics are expressed as frequencies and percentages for categorical variables and the mean±standard deviation (SD) or 95% confidence interval (95%CI) for continuous variables, with group comparisons performed using the chi-square or Fisher exact tests and 1-way ANOVA.

Competing risk analyses were conducted using Fine and Gray proportional subhazards models with death and iHF as competing events, adjusting for relevant covariates (figure S1). Cumulative incidence functions, incidence rates, and time-to-first event analyses were conducted to evaluate fatal and nonfatal cardiovascular outcomes by HF status. Medical resource use and expenditure were compared among strata using the chi-square and nonparametric tests. Poisson models were used to assess health care utilization and Gaussian models (log-costs) for expenditure. Finally, mixed-effects generalized linear models with patients as random effects were used to assess the association between HF status and health care utilization and costs, adjusting for age, sex, income, clinical complexity, time since ACS, timing of health care delivery relative to ACS, and their interaction.

All statistical tests and 95%CI were constructed with a type I error alpha level of 5%, with no adjustments for multiplicity. P-values<.05 were considered statistically significant. All analyses were performed using R software (version 4.0.2; R Foundation for Statistical Computing, Vienna, Austria).

Patient selection in this study is represented in figure S2. Between January 1, 2011 and December 31, 2021, 83 357 patients without previous HF were discharged after an ACS in Catalonia, Spain. The mean age of this overall population was 66.8±13.4 years and 71.4% (n=59 543) were men. During the study period, 19 417 (23.3%) individuals developed iHF [hereafter iHF(+) group]. Of them, 12 731 (15.3% of the entire cohort) developed iHF within the 2 years following the index event and were, therefore, considered early-onset (EO), while 6686 (8.0%) developed iHF 2 to 10 years after the index event (late onset, LO).

Table 1 compares iHF(-), iHF(EO), and iHF(LO) patients. Those with iHF were older, more often female, had more cardiovascular risk factors and prior disease, lower socioeconomic status, and less STEMI/NSTEMI. Killip II-III and anterior STEMI were more frequent in iHF(EO), while percutaneous coronary intervention (PCI) rates were lower in both iHF groups compared with the iHF(-) group (P <.001).

Fine and Gray's model identified key risk factors for developing iHF post-ACS: increasing age (subhazard ratios [sHR], per year, 1.03; 95%CI, 1.03-1.04), low income (sHR, 1.92; 95%CI, 1.45-2.54), very low income (sHR, 2.08; 95%CI, 1.56-2.78), and Killip II-III class (sHR, 2.16; 95%CI, 2.06-2.27). PCI during ACS admission was associated with lower risk (sHR, 0.72; 95%CI, 0.70–0.75). Figure S1 shows additional predictors, including hypertension (sHR, 1.12; 95%CI, 1.07-1.17), obesity (sHR, 1.16; 95%CI, 1.13-1.20), atrial fibrillation/flutter (sHR, 1.44; 95%CI, 1.39-1.49), peripheral arterial disease (sHR, 1.19; 95%CI, 1.15-1.24), and chronic kidney disease (sHR, 1.19; 95%CI, 1.14-1.23).

As shown in table 2 and figure 1A, the incidence rate of HF after ACS was 5.2 new cases per 100 person-years of follow-up (table 2 and figure 1A). Overall, 23.3% developed iHF, 21.7% of the cohort died at the end of the study period, 17.8% developed recurrent ACS, 17.4% required a new CV intervention, and 42.1% had any of the 4-point MACE.

Figure 1.

Cumulative incidence of heart failure and all-cause mortality following ACS hospitalization. A: in the whole cohort and according to B: age; C: Killip class; D: type of ACS. The shading indicates the 95% confidence interval. Competing risk between death and incident HF were considered in these analyses. ACS, acute coronary syndromes; iHF, incident heart failure; NSTEMI, non–ST-elevation myocardial infarction; STEMI, ST-segment elevation myocardial infarction.

(0.68MB).

Adverse clinical events increased markedly over time. The cumulative incidence of the 5-point MACE was 37% at 2 years, 51% at 5 years, and 66% at 10 years (figure S3). Older age and Killip class II-III were associated with higher rates of mortality and HR (figure 1). NSTEMI showed the highest incidence of death and HF, while UA surpassed STEMI in long-term HF risk. PCI rates were lower in patients with iHF.

As shown in table 2, table 3, and figure 2, iHF(+) patients experienced significantly more adverse outcomes than iHF(-) patients. Compared with iHF(-), risks were elevated for all-cause mortality (hazard ratio [HR], 3.44; 95%CI, 3.34-3.54), recurrent ACS (HR, 2.09; 95%CI, 2.02-2.16), myocardial infarction (HR, 2.73; 95%CI, 2.62-2.85), stroke/transient ischemic attack (HR, 2.09; 95%CI, 1.96-2.24), and 4-point MACE (HR, 2.75; 95%CI, 2.69–2.81).

Figure 2.

Cumulative incidence curves in patients with and without incident heart failure of A: all-cause mortality; B: new cardiovascular intervention; C: stroke or transient ischemic attack; D: new atrial fibrillation/flutter; E: 4-point major adverse cardiac event (all-cause mortality, recurrent ACS, new coronary revascularization [either surgical or percutaneous], and stroke), and F: terminal event (death, mechanic circulatory support, or heart transplantation). iHF (-), no incident heart failure; iHF (+), incident heart failure.

(0.59MB).

Health care contacts and costs increased slightly before ACS and peaked sharply during hospitalization, then declined but remained above baseline (figure 3A-3B). Mean total costs were more than twice as high in the first year (€6016±€8778) and 1.5 times higher in the second year (€4076±€7259) compared with the year before ACS (€2628±€5151), mainly driven by hospitalizations and pharmacy use. Hospital admissions were consistently higher in iHF(+) patients from 1 year before to 2 years after ACS (P <.001). Hospitalizations and pharmacy costs were the main cost drivers in both groups. Further information about trends in health care contacts and expenditure are summarized in figure 3C-D.

Figure 3.

Ratio of contacts with the health care system and mean monthly expenditure before and after an ACS in the total population (A and B), and according to iHF status (C and D). Relative risk of medical resource use (E) and expenditure (F) before and after ACS and according to iHF status across different categories of health care services. ACS, acute coronary syndromes; AP, primary care; EM, emergency medicine visits; EXT, outpatient visits; PHARM, pharmacy; HF, heart failure; HOS, hospitalizations; iHF, incident heart failure; MH, mental health visits; OTH, outpatient hospitalizations; OV, outpatient visits; PC, primary care visits; SS, social services; URG, emergency visits.

(0.79MB).

As shown in figure 3E, mixed-effects adjusted models identified iHF(+) as an independent predictor of increased health care use compared with iHF(-), including higher monthly rates of primary care visits (incidence rate ratio [IRR], 1.63; 95%CI, 1.57-1.69), total hospital admissions (IRR, 3.18; 95%CI, 2.98-3.40), unplanned admissions (IRR, 5.56; 95%CI, 5.13-6.03), emergency visits (IRR, 1.77; 95%CI, 1.67-1.87), drug prescriptions (IRR, 1.97; 95%CI, 1.90-2.05), and social/health care services (IRR, 8.19; 95%CI, 6.61-10.20). Additionally, iHF(+) was associated with significantly higher overall health care expenditure (β, 0.73; 95%CI, 0.72-0.75; P <.001) (figure 3F).

The main strengths of this study include its large cohort size (∼8 million people), long follow-up, and high-quality data from a national registry. Importantly, it is the first to analyze health care resource use and expenditure in patients with ACS and iHF.

However, several limitations must be acknowledged. First, the population-based design relied on administrative diagnostic codes, which may not fully capture HF complexity. Indeed, iHF diagnoses included both hospitalized patients and those diagnosed in primary care without hospitalization, who may differ clinically; nevertheless, this approach reflects the full real-world spectrum of HF. We lacked detailed clinical data, such as ejection fraction, natriuretic peptides, and functional class, which limited phenotyping. In addition, the diagnosis of iHF was based on criteria widely used in previous studies, combining hospitalization with a primary diagnosis of HF and primary care records validated by loop diuretic prescription. Although this approach reflects real-world practice in integrated health care systems, we acknowledge that it may introduce heterogeneity in event definition.

Second, although Catalonia's universal health care system covers> 98% of cardiac care, data from private centers or outside the region were not included. Third, socioeconomic status was assessed only by income at the index event, without accounting for education, occupation, or changes over time. Fourth, while prescription data were available, medication adherence and lifestyle changes could not be evaluated. We also lacked information on complete vs incomplete revascularization during ACS. Fifth, changes in ACS management and coding practices over the decade may have influenced results, though sensitivity analyses showed consistency. Sixth, we found women had a higher incidence of iHF; however, these results may have been influenced by differences in ACS management among centers. Finally, while our findings are likely generalizable to other universal health care systems, they may not apply to settings with different structures or financial barriers to care.