RESULTS
Baseline characteristics of the patients and comparison between the group with anemia and the control group.
A total of 13,791 patients were included, 13,454 of whom had valid Hb values and were finally included in the analysis (Figure 1). The Hb values were [(median (p25-p75)] 12 g/dL(10.6-13.3), and 7,662 (56.9%) patients had anemia. The frequency of anemia increased with age: being 46.5% in subjects under 65 years of age, 54% between 65 and 75, and 58.8% in individuals over the age of 75 (p of the linear trend < 0.001).
Table 1 shows the baseline charactistics of the population and comparison of the two groups. The mean age of the patients was 80 (10) years, and 55.5% were women. The group with anemia was older [80.6(9.5) vs. 79.2(9.5), p < 0.001), with a lower percentage of women (52.5% vs. 59.6%, p < 0.001) and a greater frequency of risk factors and cardiovascular comorbidity, except for atrial fibrillation (p = 0.33). The patients with anemia had a higher percentage of functional class for dyspnea NYHA III-IV (26.7 vs. 22.2, p < 0.001) and a Barthel index <60 points (20.7 vs. 16.3, p < 0.001). Chronic treatment included more loop and potassium sparing diuretics, beta-blockers, antiaggregants, digoxin, nitrates and statins. In relation to the complementary tests, the group with anemia showed greater renal function impairment, with a percentage of patients with a greater eGFR < 60 ml/min (70.9 vs. 53.8, p<0.001) and more hyponatremia (20.3 vs. 16.1, p<0.001). Among the biomarkers, patients with anemia had higher N-terminal brain natriuretic peptid (NTproBNP) values [8472 (13676) pg/mL] than patients without anemia [6812 (13093), p<0.001] and had a higher percentage of positive troponin values (58.4 vs. 51.1, p<0.001).
Anemia as a prognostic factor of mortality
Mortality was greater in patients with anemia: 7.5% in individuals without anemia vs. 10.7% with anemia (p < 0.001), and 21.2% vs. 31.4% (p < 0.001) at 30 days and at one year, respectively. When patients were classified based on Hb values, the mortality at 30 days and at one year was: mortality at 30 days 10.7% with Hb <7 g/dL, 14.3% with Hb between 7 and 9g/dL, 10.4% with Hb between 9 and 11g/dL, 9.5% with Hb between 11 and 13g/dL and 7.1% whit Hb >13g/dL, p-value for trend <0.001; mortality at one year 30% with Hb < 7g/dL, 36.8% with Hb between 7 and 9g/dL, 31.9% with Hb between 9 and 11g/dL, 27.6% with Hb between 11 and 13g/dL and 20% with Hb>13g/dL, p-value for trend < 0.001. Figure 2 shows the crude hazard ratios (95% confidence interval [95% CI]) for each range of Hb values of mortality at 30 days and at one year.
Patients with anemia had a greater crude in-hospital [HR(95CI%): 1.40(1.23-1.60), p<0.001], and 30-day [HR(95CI%): 1.46(1.30-1.64), p<0.001] and one-year [HR(95CI%): 1.57(1.47-1.68), p<0.001] mortality (Figure 2). In the multivariate analysis (figures 3 and 4), the effect of anemia on mortality remained significant after having controlled for confounding factors in the different partial models to the global model HR(CI95%): 1.20(1.05-1.38), p=0.009 for 30-day mortality ant 1.30(1.20-1.140), p<0.001 for one-year mortality.
The most important change in the HR (CI95%) in relation to the influence of anemia on mortality was observed when: in the case of mortality at 30 days, age and variables of comorbidity were introduced into the model, leading to a reduction to 12.3% in the HR; and in relation to mortality at one year, the change in the crude HR and that adjusted for age, sex and comorbidity was 15.2%. There were no important variations in the other partial models (Figures 3 and 4).
Data on the type of dysfunction were only available in 5427 patients (40.3%), and of these, 2936 (54%) had a reduced ejection fraction. We performed a stratified analysis of mortality at 30 days and one year between these two groups. In the case of patients with heart failure and a reduced ejection fraction, the crude HR (CI95%) was 1.43 (1.08-1.89), p=0.01 for 30-day mortality and 1.60 (1.37-1.89), p<0.001 for one-year mortality, and the adjusted values were 1.28 (0.96-1.71), p=0.09 and 1.37(1.16-1.62), p<0.001, respectively. In the case of heart failure with a preserved ejection fraction, the crude HR values were 1.52 (1.08-2.14), p=0.02 mortality at 30 days and 1.51 (1.25-1.82), p<0.001 for mortality at one year. The adjusted values were 1.48 (1.04-2.11), p=0.03 and 1.35 (1.11-1.64), p=0.003, respectively.
Effect of anemia on mortality in the individual time periods studied.
In the analysis of risk for the individual time periods (Figure 5), the effect on mortality was lost in the intermediate follow-ups but gained increasing importance after the eighth month of follow-up, being greater in patients who were alive in the last month of follow-up.