Anaemia, erythrocytosis, and empagliflozin in heart failure with preserved ejection fraction: the EMPEROR-Preserved trial

Many patients with heart failure (HF) and a preserved ejection fraction have anaemia as a consequence of ‘chronic disease’. Anaemia worsens HF symptoms and has been associated with poor cardiovascular outcomes.1 Sodium-glucose cotransporter 2 inhibitors (SGLT2i) increase haemoglobin (and haematocrit), often leading to the correction of anaemia and, in some patients, SGLT2i use may lead to erythrocytosis.2–5 While the association of anaemia with poor outcomes has been documented in HF, the association of erythrocytosis with outcomes in HF is less well established.5 The association of anaemia and erythrocytosis with clinical outcomes, the impact of SGLT2i on anaemia and erythrocytosis throughout the follow-up, and the effect of SGLT2i on cardiovascular and kidney outcomes in patients with and without anaemia or erythrocytosis are explored herein using the EMPEROR-Preserved (NCT03057951) data. The design and primary results of the EMPEROR-Preserved trial have been published previously.6 The median follow-up was 26.2 months. The primary endpoint was a composite of cardiovascular death or hospitalization for HF. Additionally, the individual components of the primary outcome; a kidney composite outcome of sustained ≥40% estimated glomerular filtration rate decline from baseline, dialysis, transplantation, or end-stage renal disease; a composite of myocardial infarction, stroke, or cardiovascular death; deep vein thrombosis or pulmonary thromboembolism; and all-cause mortality were also studied. Anaemia and erythrocytosis were classified according to the World Health Organization definition of haemoglobin level < 12.0 g/dL in women and <13.0 g/dL in men, and haemoglobin level > 16.0 g/dL in women and >16.5 g/dL in men, respectively. Haemoglobin (and haematocrit) was obtained from the laboratory results at the main study visits (i.e. baseline, Weeks 4, 12, 32, and 52, and every 24 weeks thereafter). The effect of empagliflozin (vs. placebo) on haemoglobin (and haematocrit) over time was studied using a mixed model for random effects with a treatment-by-visit interaction, and random effects at the patient level, adjusting on previously specified variables.6 All analyses were performed using Stata®, v18. All analyses were performed within the Vivli® (https://vivli.org/) platform with the Vivli ID: 00009383. The trial sponsor was not involved in this analysis. Of the 5988 patients enrolled in the trial, at baseline, 1703 (28.4%) had anaemia. The mean haemoglobin (haematocrit) level in patients with anaemia was 11.6 ± .9 g/dL (35.8 ± 2.9%) and in patients without anaemia was 14.0 ± 1.2 (42.9 ± 3.7%). Compared to patients without anaemia, those with anaemia were older, more frequently men, with diabetes, HF of ischaemic aetiology, poorer renal function, and higher N-terminal pro-B-type natriuretic peptide levels. At baseline, 127 (2.1%) patients had erythrocytosis. The mean haemoglobin (haematocrit) level in patients with erythrocytosis was 17.2 ± .8 g/dL (52.6 ± 3.1%) and in patients without erythrocytosis was 13.2 ± 1.5 g/dL (40.7 ± 4.5%). Compared to patients without erythrocytosis, those with erythrocytosis were younger, more often smokers, with pulmonary disease. In the placebo group, the presence of anaemia at baseline was independently associated with an increased risk (1.4× to 2.0×) of occurrence of all studied outcomes (see Methods section) compared to patients without anaemia. The adjusted hazard ratio (HR) and 95% confidence interval (CI) for the primary outcome were 1.66 (1.38–1.99; P < .001). The presence of erythrocytosis at baseline was not associated with any of the studied outcomes compared to patients without erythrocytosis. The adjusted HR (95% CI) for the primary outcome was .90 (.51–1.61; P = .73). The occurrence of erythrocytosis during the follow-up (in time-updated models) was not associated with atherothrombotic events. The presence of anaemia did not modify the effect of empagliflozin on the studied outcomes (all interaction P > .1). The empagliflozin vs. placebo HR (95% CI) for the primary outcome in patients with anaemia was .83 (.68–1.01) and in patients without anaemia was .77 (.64–.91; interaction P = .56). Despite a similar relative reduction in events with empagliflozin (vs. placebo), the absolute risk reduction was greater among patients with anaemia at baseline who had a higher absolute risk of events (Figure 1A). Anaemia and erythrocytosis in EMPEROR-Preserved. (A) Primary outcome occurrence according to anaemia status and treatment group. (B) Effect of empagliflozin on haemoglobin levels over time. (C) Effect of empagliflozin on haemoglobin levels over time among patients with and without CKD. (D) Effect of empagliflozin on anaemia and erythrocytosis over time. (A) Cumulative incidence of primary outcome events expressed in %. Despite a higher risk of primary outcome events among patients with anaemia, the relative reduction of events with empagliflozin (vs. placebo) was similar regardless of anaemia status. (B) Throughout the follow-up, the average (standard error) haemoglobin level in the placebo group was 13.24 (.01) g/dL and 13.79 (.01) g/dL in the empagliflozin group, with a between-group (empagliflozin vs. placebo) difference of .55 (.02) g/dL (joint P < .001). Regarding specific time-points, at baseline the between-group difference was .01 (.03) g/dL, at Week 12 was .58 (.03) g/dL, at Week 52 was .74 (.03) g/dL, and at Week 100 was .82 (.03) g/dL. (C) Among patients without CKD, throughout the follow-up, the average (standard error) haemoglobin level in the placebo group was 13.29 (.02) g/dL and 13.85 (.02) g/dL in the empagliflozin group, with a between-group (empagliflozin vs. placebo) difference of .56 (.03) g/dL (joint P < .001). Regarding specific time-points, at baseline the between-group difference was .00 (.04) g/dL, at Week 12 was .62 (.04) g/dL, at Week 52 was .78 (.04) g/dL, and at Week 100 was .81 (.04) g/dL. Among patients with CKD, throughout the follow-up, the average (standard error) haemoglobin level in the placebo group was 13.18 (.02) g/dL and 13.72 (.02) g/dL in the empagliflozin group, with a between-group (empagliflozin vs. placebo) difference of .54 (.03) g/dL (joint P < .001). Regarding specific time-points, at baseline the between-group difference was .02 (.04) g/dL, at Week 12 was .55 (.04) g/dL, at Week 52 was .70 (.04) g/dL, and at Week 100 was .83 (.05) g/dL. No CKD vs. CKD interaction P = .69. (D) Anaemia: at specific time-points, the corresponding anaemia n/N (%) and odds ratios were: baseline: placebo 852/2989 (28.5) vs. empagliflozin 848/2998 (28.3), OR .99, 95% CI .88–1.11, P = .85. Week 12: placebo 803/2767 (29.0) vs. empagliflozin 552/2787 (19.8), OR .60, 95% CI .53–.68, P < .001. Week 52: placebo 661/2329 (28.4) vs. empagliflozin 433/2332 (18.6), OR .58, 95% CI .50–.66, P < .001. Week 100: placebo 448/1445 (31.0) vs. empagliflozin 237/1434 (16.5), OR .44, 95% CI .37–.53, P < .001. Erythrocytosis: at specific time-points, the corresponding erythrocytosis n/N (%) and odds ratios were: baseline: placebo 79/2989 (2.6) vs. empagliflozin 47/2998 (1.6), OR .60, 95% CI .42–.86, P = .006. Week 12: placebo 64/2767 (2.3) vs. empagliflozin 149/2787 (5.3), OR 2.36, 95% CI 1.78–3.14, P < .001. Week 52: placebo 51/2329 (2.2) vs. empagliflozin 160/2332 (6.9), OR 3.29, 95% CI 2.39–4.53, P < .001. Week 100: placebo 32/1445 (2.2) vs. empagliflozin 112/1434 (7.8), OR 3.74, 95% CI 2.51–5.58, P < .001. Empa., empagliflozin; Pbo., placebo; SE, standard error; HF, heart failure; CKD, chronic kidney disease; OR, odds ratio. *P < .001 The presence of erythrocytosis also did not modify the effect of empagliflozin on the studied outcomes (all interaction P > .05). However, despite the small number of events (24 events in total), a tendency to an increased risk of primary outcome events with empagliflozin was observed among patients with erythrocytosis at baseline: no erythrocytosis HR (95% CI) .78 (.68–.89) and erythrocytosis 1.65 (.74–3.68; interaction P = .069). Throughout the follow-up, the average (standard error) haemoglobin level in the placebo group was 13.24 (.01) g/dL and 13.79 (.01) g/dL in the empagliflozin group, with a between-group (empagliflozin vs. placebo) difference of .55 (.02) g/dL (joint P < .001; Figure 1B). The corresponding average (standard error) haematocrit level increase with empagliflozin (vs. placebo) was 1.94 (.06) % (joint P < .001). The presence of chronic kidney disease did not modify the effect of empagliflozin on haemoglobin levels (interaction P = .69; Figure 1C). Empagliflozin (vs. placebo) reduced the odds of experiencing anaemia by 40% at Week 52: odds ratio (OR) .58 (95% CI .50–.66; P < .001). On the other hand, empagliflozin (vs. placebo) increased the odds of experiencing erythrocytosis by 3.3× at Week 52 (OR 3.29, 95% CI 2.39–4.53; P < .001; Figure 1D). An increase in haemoglobin levels was associated with higher odds of meaningful (≥5 points) Kansas City Cardiomyopathy Questionnaire clinical summary score improvement (and lower odds of deterioration) among patients randomized to empagliflozin but not among those randomized to placebo. The present study adds novel and clinically relevant information: (i) the treatment effect of empagliflozin was not influenced by anaemia status and empagliflozin led to an important reduction of anaemia throughout the follow-up; (ii) increment in haemoglobin and anaemia correction with empagliflozin was of similar magnitude regardless of chronic kidney disease; and (iii) empagliflozin increased the proportion of patients with erythrocytosis throughout the follow-up, but erythrocytosis was not associated with atherothrombotic events. Still, among patients with erythrocytosis at baseline, empagliflozin use was associated with a tendency towards an increased risk of primary outcome events, a finding limited by the small number of events and by the small number of patients with erythrocytosis.7,8 In conclusion, empagliflozin increased haemoglobin levels and substantially reduced the incidence of anaemia. While there was an increase in erythrocytosis, this was not associated with atherothrombotic events. Further studies should clarify the risk factors for the occurrence of erythrocytosis with SGLT2i and whether there is an erythrocytosis threshold above which the risk of atherothrombotic events may be increased. This publication is based on research using data from data contributors Boehringer Ingelheim that has been made available through Vivli, Inc. Vivli has not contributed to or approved, and is not in any way responsible for, the contents of this publication. J.P.F. reports consulting fees from Boehringer Ingelheim during the conduct of the study, and research support from Boehringer Ingelheim, AstraZeneca, Novartis, Bayer, Abbott, Salamandra, and Bial outside. F.Z. reports personal fees from Boehringer Ingelheim during the conduct of the study; personal fees from Janssen, Novartis, Boston Scientific, Amgen, CVRx, AstraZeneca, Vifor Fresenius, Cardior, Cereno pharmaceutical, Applied Therapeutics, Merck, Bayer, and Cellprothera outside the submitted work; and other support from CVCT and Cardiorenal, outside the submitted work. S.J.P. reports personal fees from Boehringer Ingelheim during the conduct of the study. J.B. reports consultancy fees from Boehringer Ingelheim during the conduct of the study; and consultancy fees from Abbott, Adrenomed, Amgen, Applied Therapeutics, Array, AstraZeneca, Bayer, BerlinCures, Boehringer Ingelheim, Cardior, CVRx, Foundry, G3 Pharma, Imbria, Impulse Dynamics, Innolife, Janssen, LivaNova, Luitpold, Medtronic, Merck, Novartis, NovoNordisk, Relypsa, Roche, Sanofi, Sequana Medical, V-Wave, and Vifor, outside the submitted work. G.F. reports receiving payment from Boehringer Ingelheim for being a trial committee member during the conduct of the study and from Medtronic, Vifor, Servier, and Novartis for being a trial committee member outside the submitted work. M.P. reports personal fees from Abbvie, Actavis, Amarin, Amgen, AstraZeneca, Boehringer Ingelheim, Caladrius, Casana, CSL Behring, Cytokinetics, Imara, Lilly, Moderna, Novartis, Reata, Relypsa, and Salamandras. S.D.A. has received research support from Abbott Vascular and Vifor International and personal fees from Abbott Vascular, Bayer, Boehringer Ingelheim, BRAHMS, Cardiac Dimensions, Impulse Dynamics, Novartis, Servier, and Vifor Pharma. F.V.-N. has received consulting or speaker fees from AstraZeneca, Bayer, Daiichi Sankyo, and Ultragenyx. Data are available through the Vivli® platform. The EMPEROR-Preserved trial was funded by Boehringer Ingelheim and Eli Lilly (EMPEROR-Preserved ClinicalTrials.gov number, NCT03057951). The protocol was approved by the Ethical Committee of each of the 622 participating sites in 23 countries, and all patients gave written informed consent to participate in the study. EMPEROR-Preserved ClinicalTrials.gov number, NCT03057951.