ABC | Volume 114, Nº2, February 2020

Original Article Fernandes et al. Heart failure trends in Less developed Brazil Arq Bras Cardiol. 2020; 114(2):222-231 Study population The population of interest was composed by Brazilians older than 15 years that used any healthcare services under the primary diagnosis of HF, represented by the code I50 of the International Classification of Diseases 10 th Revision (ICD‑10), between 2008 and 2017. Variables Epidemiological data on HF were extracted, including absolute and relative mortality of the population, in-hospital mortality (absolute numbers), in-hospital mortality rate, number of hospital admissions and length of hospitalization. Variables were stratified by year, gender and age groups (15-19, 20-29, 30-39, 40-49, 50-59, 60-69, 70-79 and≥80 years). In-hospital data from the period of 2008 to 2017, and population data from 2008 to 2015 were available. The last population census conducted by the Brazilian Institute of Geography and Statistics (IBGE) 5 in 2010 was also used. Data analysis Categorical variables were expressed as frequencies and continuous variables as mean ± standard deviation (SD). In-hospital mortality rate fromHF was obtained by dividing the number of all in-hospital HF deaths in Paraiba or Brazil by the number of hospitalizations for HF in the corresponding year. Population mortality rate from HF was calculated by dividing the number of all HF deaths in Paraiba or Brazil by the respective population in the corresponding year. The Statistical Package for the Social Sciences (SPSS) version 21.0 (SPPS Inc., Chicago, USA) was used for the analysis. We used the Shapiro-Wilk’s test to test the normality of data distribution for further analyses. The Pearson correlation was used to evaluate the correlation between numerical variables with normal distribution. The Chi‑square test was performed using a contingency table and the linear‑by-linear association test, also known as Mantel-Haenzsel test for trends, which is equivalent to the Cochran-Armitage test for trends available in other statistical packages. 9 The level of significance was set at 5%. Results Descriptive statistics of our variables are presented in Table 1. Hospitalizations The total number of HF admissions in Paraiba state between 2008 and 2017 was 51,172, representing the leading cause of hospitalizations due to cardiovascular diseases (29.4%), followed by other ischemic diseases of the heart (13%), stroke (11%), primary hypertension (10%) and acutemyocardial infarction (5%). During the same period, HF was also the leading cardiovascular cause of hospitalization in Brazil, with 2,380,133 cases (21%). HF was responsible for 2.54% and 2.25% of all causes of hospitalization in Paraiba and in Brazil, respectively. A downward trend in the absolute number of hospitalizations from HF in Paraiba and Brazil was observed between 2008 and 2017, corresponding to a decrease of 62% (R = -0.970; p = 0.004; Table 2; Figure 1A) and 34% (R = -0.964; p = 0.004; Table 3; Figure 1B), respectively. The frequency of males hospitalized for HF was 52% in Paraiba and 51% in Brazil. When stratified by age, individuals older than 60 years old corresponded to 71% and 73% of all the cases of HF admissions in Paraiba and Brazil, respectively, with the highest frequency in the age range from 70 to 79 years old. Absolute mortality of population The absolute mortality from HF of the population showed a non-significant decline from2008 to 2015 in Paraiba (R=-0.513; p = 0.175; Table 2) and Brazil (R= -0.412; p = 0.276; Table 3), with no difference by gender.Women represented 53%of deaths in Paraiba and 52% in Brazil. In Paraiba, the decrease in absolute deaths from HF in the population across all age categories was not statistically significant (Table 2). Between 2008 and 2015, the highest proportion of deaths from HF occurred at the age group of ≥ 80 years old in both men and women in Paraiba (50% and 59%, respectively) and in Brazil (38% and 52%, respectively). The proportions of deaths from HF at the age ≥ 60 years old in Paraiba was 87% in men and 90% in women and, in Brazil, 83% in men and 89% in women. Population mortality rate The mean mortality rate from HF in the population was 19.2/100,000 (±1.09) in Paraiba and 14.0/100,000 (±0.53) in Brazil, with a significant decline of 10.7% (R = -0.751; p = 0.047; Table 2) in Paraiba and 7.7% (R = -0.905; p = 0.017; Table 3) in Brazil between 2008 and 2015, respectively (Figure 2). Absolute in-hospital mortality The absolute in-hospital HF mortality, between 2008 and 2017, showed a significant decrease of 37.5% in Paraiba (R = -0.824; p = 0.013; Table 2; Figure 3B) and a non‑significant 14.6% decrease in Brazil (R = -0.504; p = 0.131; Table 3; Figure 3B). In the stratified analysis, a significant decrease in the absolute in-hospital deaths from HF was observed for both men and women in Paraiba (R = -0.837; p = 0.012 and R = -0.762; p = 0.022; Table 2); this statistically significant trend by sex was not observed in Brazil (Table 3). Individuals older than 80 years old presented the highest proportion of absolute in-hospital HF deaths in Paraiba and Brazil, from 2008 to 2017, (37% and 32%, respectively) (Figure 4). In Paraiba, there was a statistically significant reduction in in-hospital deaths from HF for the age categories: 20-29 years (p = 0.010), 30-39 years (p = 0.008), 40-49 years (p = 0.029), 50-59 years (p = 0.025) and 70-79 years (p = 0.009) (Table 2). Further data on the absolute number of in-hospital deaths from HF per age range in Brazil are specified in Table 3. In-hospital mortality rate The in-hospital HF mortality rate increased significantly by 65.1% in Paraiba (R = 0.917; p = 0.006; Table 2), from 6.6% 223