Methods

Data sources

Data from five Swedish registries were used for this study, interlinked via the unique personal identification number held by all Swedish residents. The study population was derived from the Swedish HF Registry (SwedeHF) which has been previously described.12 The inclusion criterion is clinician-judged HF, since 2017 defined by International Classification of Diseases, 10th Revision (ICD-10) codes reported in Supplementary data online, Table S1. In 2023, SwedeHF covered 59% of prevalent HF cases in Sweden.13 Pharmacy dispensations of prescribed drugs were extracted from the National Prescribed Drugs Register (NPDR), which is managed by the National Board of Health and Welfare, and collects information on all prescribed drugs [categorized according to Anatomical Therapeutic Chemical (ATC) codes] dispensed in Swedish pharmacies since July 2005. Socioeconomic variables were obtained through Statistics Sweden’s Longitudinal Integrated Database for Health Insurance and Labour Market Studies and the Register of the Total Population. The National Patient Register and Cause of Death Register, both managed by the National Board of Health and Welfare, provided additional comorbidity data, HHF, and cause-specific mortality. The cross-registry linkage and current analysis were approved by the Swedish Ethical Review Authority.

Patients

Patients with HFrEF enrolled in SwedeHF between 1 January 2016 and 31 December 2023 were included. The index date was defined as the day of the outpatient visit or hospital discharge prompting the latest entry in SwedeHF. Heart failure with reduced ejection fraction was defined as ejection fraction (EF) < 40% rather than ≤40%, due to the categorization of SwedeHF-reported EF in most patients. Follow-up was censored at 3 years, 31 December 2023, death or move from Sweden. The patient selection flow chart is presented in Supplementary data online, Table S2.

Variable definitions

Prescribed treatments at index were identified by pharmacy dispensations within 4 months prior until 14 days after the index date and assessed according to drug class and number of GDMT components (i.e. single, dual, triple, or quadruple therapy).

Prescribed daily dosages were calculated by multiplying the strength of the prescribed tablet with the number of pills per day prescribed, as algorithmically extracted from free text dose instructions reported in NPDR. For cases where prescribed dosages were missing or uninterpretable, the dose was imputed as the mean of non-missing dosages for the same ATC class. The daily dosage prescribed at the baseline dispensation was further quantified as the percentage of the substance-specific target dose recommended in the 2021 European Society of Cardiology HF guidelines (see Supplementary data online, Table S3).1

Proportion of days covered (PDC) was calculated for each dispensation interval (per medication class) by dividing the number of days covered (calculated as dispensed total dose divided by daily prescribed dosage) with the number of days between dispensations. Good adherence for each prescribed regimen was defined as ≥80% PDC during the first year post-index in the main analysis. The ≥80% PDC threshold was selected as it is well-established in pharmacoepidemiological studies.14 A sensitivity analysis was performed denoting good adherence instead as ≥88% PDC, based on a prior small study reporting this as the most prognostic cut-off in patients with HF.15

Patients filling prescriptions at 6 ± 2, 12 ± 2, 18 ± 2, and 24 ± 2 months post-index were considered persistent at 0.5, 1, 1.5, and 2 years, respectively. A 4-month time window (i.e. ± 2 months) was selected to add 1 month grace period (accounting for potential pill stockpiling) to the ∼3 months covered by the most commonly prescribed pill packages. Variable definitions, including ATC codes for prescribed treatments and ICD-10 codes for comorbidity variables and cause-specific outcomes, are reported in Supplementary data online, Table S4.

Statistical analysis

Baseline characteristics were presented as median [interquartile range (IQR)] if continuous, and frequency (per cent) if categorical, and compared according to number of GDMT components at baseline by Kruskal–Wallis and _χ_2 test, respectively.

The proportion of patients being prescribed each therapy regimen was calculated overall and according to index year. For calculating the proportion of patients with good adherence and persistence, the denominator was the number of patients using the respective regimen at index (including prevalent as well as first-time users). For persistence at 0.5, 1, 1.5, and 2 years, patients also had to be alive at 6 + 2, 12 + 2, 18 + 2, and 24 + 2 months post-index. A sensitivity analysis was performed separately in first-time users, defined as patients using the respective regimen at index but without a prior prescription dispensation for that regimen in the past 12 months.

Patient characteristics associated with good adherence and 1-year persistence were assessed by multivariable logistic regression including variables reported in Table 1, with results reported as odds ratios (OR) with 95% confidence intervals (CI).

Variable | Overall (n = 35 215) | None (n = 471, 1%) | Mono (n = 2093, 6%) | Double (n = 11 205, 32%) | Triple (n = 14 718, 42%) | Quadruple (n = 6728, 19%) | P-value | Missing (%)
Demographic characteristics |  |  |  |  |  | 
Index yeara | 2020 [2018–2022] | 2019 [2017–2021] | 2018 [2017–2020] | 2019 [2017–2021] | 2019 [2018–2021] | 2023 [2022–2023] | <.001 | 0
Index year |  |  |  |  |  |  | <.001 | 0
2016–20 | 19 966 (57) | 348 (74) | 1619 (77) | 8403 (75) | 9596 (65) | <10 |  | 
2021–23 | 15 249 (43) | 123 (26) | 474 (23) | 2802 (25) | 5122 (35) | 6728 (100) |  | 
Sexa |  |  |  |  |  |  | <.001 | 0
Female | 9891 (28) | 121 (26) | 594 (28) | 3346 (30) | 4046 (27) | 1784 (27) |  | 
Male | 25 324 (72) | 350 (74) | 1499 (72) | 7859 (70) | 10 672 (73) | 4944 (73) |  | 
Age (years) | 74 [64–80] | 73 [61–80] | 78 [69–85] | 75 [66–82] | 73 [64–79] | 72 [62–78] | <.001 | 0
Age (years)a |  |  |  |  |  | <.001 | 0
<70 | 13 053 (37) | 208 (44) | 524 (25) | 3620 (32) | 5737 (39) | 2964 (44) |  | 
70–80 | 13 785 (39) | 148 (31) | 733 (35) | 4300 (38) | 5917 (40) | 2687 (40) |  | 
>80 | 8377 (24) | 115 (24) | 836 (40) | 3285 (29) | 3064 (21) | 1077 (16) |  | 
Socioeconomic characteristics |  |  |  |  |  | 
Family situationa |  |  |  |  |  | <.001 | 0
Cohabitating | 19 033 (54) | 194 (42) | 1060 (51) | 5944 (53) | 8158 (56) | 3677 (55) |  | 
Living alone | 16 139 (46) | 271 (58) | 1031 (49) | 5252 (47) | 6541 (44) | 3044 (45) |  | 
Marital status |  |  |  |  |  | <.001 | 0
Married | 17 111 (49) | 173 (37) | 989 (47) | 5441 (49) | 7276 (49) | 3232 (48) |  | 
Single/widowed/divorced | 18 061 (51) | 292 (63) | 1102 (53) | 5755 (51) | 7423 (51) | 3489 (52) |  | 
Childrena | 29 046 (82) | 354 (75) | 1755 (84) | 9346 (83) | 12 137 (82) | 5454 (81) | <.001 | 0
Educationa |  |  |  |  |  |  | <.001 | 1
Compulsory school | 11 610 (33) | 153 (34) | 813 (40) | 4048 (37) | 4837 (33) | 1759 (26) |  | 
Secondary school | 15 621 (45) | 209 (46) | 824 (40) | 4807 (43) | 6541 (45) | 3240 (49) |  | 
University | 7500 (22) | 93 (20) | 411 (20) | 2210 (20) | 3138 (22) | 1648 (25) |  | 
Incomea |  |  |  |  |  |  | <.001 | 0
1st tertile within year | 11 598 (33) | 205 (44) | 738 (35) | 3748 (33) | 4753 (32) | 2154 (32) |  | 
2nd tertile within year | 11 604 (33) | 158 (34) | 749 (36) | 3858 (34) | 4720 (32) | 2119 (32) |  | 
3rd tertile within year | 11 970 (34) | 102 (22) | 604 (29) | 3590 (32) | 5226 (36) | 2448 (36) |  | 
Clinical characteristics |  |  |  |  |  |  | 
Duration of HF (months)a |  |  |  |  | <.001 | 2
<6 | 16 553 (48) | 153 (34) | 786 (39) | 5482 (50) | 6496 (45) | 3636 (55) |  | 
≥6 | 17 882 (52) | 296 (66) | 1243 (61) | 5410 (50) | 7923 (55) | 3010 (45) |  | 
Previous HHF within 3 monthsa | 11 325 (32) | 170 (36) | 632 (30) | 3359 (30) | 4991 (34) | 2173 (32) | <.001 | 0
Previous HHF within 1 year | 14 637 (42) | 215 (46) | 797 (38) | 4196 (37) | 6551 (45) | 2878 (43) | <.001 | 0
Follow-up referral HF nurse clinica | 29 810 (88) | 291 (71) | 1515 (78) | 9200 (86) | 12 583 (89) | 6221 (95) | <.001 | 4
Follow-up referral specialitya |  |  |  |  | <.001 | 2
Primary care/other | 5208 (15) | 140 (32) | 656 (33) | 2163 (20) | 1787 (12) | 462 (7) |  | 
Hospital | 29 203 (85) | 297 (68) | 1338 (67) | 8758 (80) | 12 642 (88) | 6168 (93) |  | 
LVEF, n (%)a |  |  |  |  |  |  | <.001 | 0
30%–39% | 21 037 (60) | 261 (55) | 1427 (68) | 7156 (64) | 8433 (57) | 3760 (56) |  | 
<30% | 14 178 (40) | 210 (45) | 666 (32) | 4049 (36) | 6285 (43) | 2968 (44) |  | 
NYHA class |  |  |  |  |  |  | <.001 | 18
I | 3281 (11) | 31 (10) | 153 (10) | 998 (11) | 1315 (11) | 784 (13) |  | 
II | 14 604 (50) | 125 (42) | 671 (46) | 4436 (50) | 6200 (50) | 3172 (53) |  | 
III | 10 706 (37) | 127 (42) | 606 (41) | 3364 (38) | 4660 (38) | 1949 (33) |  | 
IV | 410 (1) | 18 (6) | 40 (3) | 134 (2) | 172 (1) | 46 (1) |  | 
NYHA classa |  |  |  |  |  | <.001 | 18
I–II | 17 885 (62) | 156 (52) | 824 (56) | 5434 (61) | 7515 (61) | 3956 (66) |  | 
III–IV | 11 116 (38) | 145 (48) | 646 (44) | 3498 (39) | 4832 (39) | 1995 (34) |  | 
BMI (kg/m2) | 26 [24–30] | 26 [23–29] | 26 [23–29] | 26 [23–30] | 27 [24–30] | 27 [24–31] | <.001 | 22
BMI (kg/m2)a |  |  |  |  |  | <.001 | 22
<30 | 20 321 (74) | 258 (78) | 1213 (78) | 6662 (76) | 8456 (73) | 3732 (71) |  | 
≥30 | 7182 (26) | 72 (22) | 350 (22) | 2147 (24) | 3120 (27) | 1493 (29) |  | 
Systolic blood pressure (mmHg) | 120 [110–135] | 122 [107–140] | 125 [112–140] | 123 [110–139] | 120 [110–135] | 120 [105–130] | <.001 | 4
Systolic blood pressure (mmHg)a |  |  |  |  | <.001 | 4
100–139 | 23 582 (70) | 280 (63) | 1335 (67) | 7379 (68) | 9994 (70) | 4594 (71) |  | 
<100 | 3031 (9) | 50 (11) | 119 (6) | 760 (7) | 1340 (9) | 762 (12) |  | 
≥140 | 7279 (21) | 114 (26) | 551 (27) | 2679 (25) | 2857 (20) | 1078 (17) |  | 
Diastolic blood pressure (mmHg) | 72 [65–80] | 74 [65–81] | 74 [65–80] | 73 [65–80] | 71 [65–80] | 71 [65–80] | <.001 | 4
Mean arterial pressure (mmHg) | 89 [81–98] | 90 [80–100] | 91 [82–100] | 90 [82–100] | 88 [80–97] | 88 [80–97] | <.001 | 4
Heart rate (b.p.m.) | 71 [62–81] | 78 [68–90] | 73 [65–84] | 72 [63–82] | 70 [62–80] | 70 [61–80] | <.001 | 4
Heart rate (b.p.m.)a |  |  |  |  |  | <.001 | 4
50–70 | 15 925 (47) | 139 (32) | 818 (41) | 4915 (46) | 6838 (48) | 3215 (50) |  | 
<50 | 751 (2) | 11 (2) | 36 (2) | 198 (2) | 334 (2) | 172 (3) |  | 
>70 | 17 020 (51) | 290 (66) | 1125 (57) | 5644 (52) | 6943 (49) | 3018 (47) |  | 
EQ-5D-VAS | 70 [50–80] | 70 [50–80] | 65 [50–80] | 70 [50–80] | 70 [50–80] | 70 [50–80] | .028 | 45
EQ-5D-VAS |  |  |  |  |  |  | .035 | 45
0–25 | 868 (4) | <10 | 43 (5) | 276 (5) | 370 (4) | 170 (4) |  | 
26–50 | 4710 (24) | 38 (27) | 235 (29) | 1474 (25) | 1964 (24) | 999 (24) |  | 
51–75 | 8009 (41) | 51 (36) | 312 (38) | 2389 (41) | 3476 (42) | 1781 (42) |  | 
76–100 | 5780 (30) | 44 (31) | 230 (28) | 1758 (30) | 2448 (30) | 1300 (31) |  | 
Laboratory characteristics |  |  |  |  |  | 
eGFR (mL/min/1.73 m2) | 69 [52–88] | 65 [43–90] | 59 [38–81] | 67 [48–87] | 70 [54–88] | 73 [57–89] | <.001 | 4
eGFR (mL/min/1.73 m2)a |  |  |  |  |  | <.001 | 4
≥60 | 21 616 (64) | 254 (57) | 961 (48) | 6433 (60) | 9403 (66) | 4565 (70) |  | 
30–59 | 10 749 (32) | 131 (30) | 718 (36) | 3591 (33) | 4454 (31) | 1855 (28) |  | 
<30 | 1611 (5) | 58 (13) | 318 (16) | 749 (7) | 388 (3) | 98 (2) |  | 
Potassium (mmol/L) | 4.3 [4.0–4.5] | 4.2 [3.9–4.4] | 4.2 [3.9–4.5] | 4.2 [4.0–4.5] | 4.3 [4.0–4.6] | 4.3 [4.0–4.6] | <.001 | 4
Potassium (mmol/L)a |  |  |  |  |  | <.001 | 4
3.5–5 (normokalaemia) | 31 616 (94) | 395 (90) | 1841 (93) | 10 049 (94) | 13 196 (93) | 6135 (95) |  | 
<3.5 (hypokalaemia) | 867 (3) | 25 (6) | 76 (4) | 281 (3) | 354 (2) | 131 (2) |  | 
>5 (hyperkalaemia) | 1317 (4) | 18 (4) | 73 (4) | 374 (3) | 631 (4) | 221 (3) |  | 
Haemoglobin (g/L) | 137 [124–149] | 130 [116–145] | 130 [116–143] | 135 [121–147] | 137 [125–148] | 142 [130–154] | <.001 | 13
NT-proBNP (pg/mL) | 2010 [813–4744] | 3220 [1178–8006] | 2803 [1077–7905] | 2364 [956–5600] | 1925 [798–4428] | 1548 [648–3590] | <.001 | 19
NT-proBNP (pg/mL)a |  |  |  |  |  | <.001 | 19
1st tertile | 9457 (33) | 91 (24) | 418 (26) | 2505 (29) | 4113 (34) | 2330 (40) |  | 
2nd tertile | 9457 (33) | 102 (27) | 473 (29) | 2764 (32) | 4112 (34) | 2006 (34) |  | 
3rd tertile | 9744 (34) | 179 (48) | 718 (45) | 3390 (39) | 3914 (32) | 1543 (26) |  | 
Comorbidities |  |  |  |  |  |  | 
Smokinga | 2978 (11) | 52 (20) | 135 (9) | 952 (11) | 1262 (11) | 577 (11) | <.001 | 23
Diabetesa | 7780 (23) | 98 (22) | 472 (24) | 2302 (21) | 3246 (23) | 1662 (25) | <.001 | 3
Diabetes type |  |  |  |  |  | <.001 | 3
No | 26 454 (77) | 352 (78) | 1533 (77) | 8503 (79) | 11 066 (77) | 5000 (75) |  | 
Type I | 460 (1) | <10 | 32 (2) | 172 (2) | 206 (1) | 43 (1) |  | 
Type II | 7298 (21) | 91 (20) | 438 (22) | 2126 (20) | 3027 (21) | 1616 (24) |  | 
Hypertensiona | 23 364 (66) | 299 (63) | 1495 (71) | 7293 (65) | 9873 (67) | 4404 (65) | <.001 | 0
Ischaemic heart diseasea | 17 900 (51) | 249 (53) | 1156 (55) | 5683 (51) | 7668 (52) | 3144 (47) | <.001 | 0
Myocardial infarction | 13 280 (38) | 201 (43) | 868 (41) | 4227 (38) | 5661 (38) | 2323 (35) | <.001 | 0
Revascularization | 10 676 (31) | 133 (30) | 637 (32) | 3278 (30) | 4670 (32) | 1958 (29) | <.001 | 2
Dilated cardiomyopathy | 5217 (15) | 72 (15) | 181 (9) | 1208 (11) | 2646 (18) | 1110 (16) | <.001 | 0
Peripheral artery diseasea | 3155 (9) | 49 (10) | 276 (13) | 1021 (9) | 1284 (9) | 525 (8) | <.001 | 0
Atrial fibrillation/fluttera | 18 921 (54) | 214 (45) | 1164 (56) | 6173 (55) | 7971 (54) | 3399 (51) | <.001 | 0
Strokea | 4505 (13) | 91 (19) | 372 (18) | 1484 (13) | 1779 (12) | 779 (12) | <.001 | 0
TIA | 1021 (3) | 12 (3) | 71 (3) | 371 (3) | 396 (3) | 171 (3) | .007 | 0
Anaemiaa | 8753 (29) | 182 (43) | 779 (42) | 3153 (32) | 3522 (28) | 1117 (19) | <.001 | 13
Valvular diseasea | 6142 (17) | 101 (21) | 454 (22) | 1902 (17) | 2616 (18) | 1069 (16) | <.001 | 0
Hyperkalaemia | 735 (2) | 14 (3) | 67 (3) | 274 (2) | 284 (2) | 96 (1) | <.001 | 0
Dialysis | 372 (1) | 15 (3) | 63 (3) | 159 (1) | 103 (1) | 32 (0) | <.001 | 0
Liver disease | 900 (3) | 36 (8) | 42 (2) | 253 (2) | 406 (3) | 163 (2) | <.001 | 0
COPDa | 3950 (11) | 55 (12) | 269 (13) | 1357 (12) | 1640 (11) | 629 (9) | <.001 | 0
Malignant cancer within 3 yearsa | 4460 (13) | 75 (16) | 365 (17) | 1565 (14) | 1713 (12) | 742 (11) | <.001 | 0
Musculoskeletal/connective tissue diseases within 3 yearsa | 10 263 (29) | 153 (32) | 752 (36) | 3481 (31) | 4118 (28) | 1759 (26) | <.001 | 0
Alcohol abusea | 1424 (4) | 33 (7) | 67 (3) | 415 (4) | 598 (4) | 311 (5) | <.001 | 0
Charlson Comorbidity Index | 2 [1–4] | 3 [2–5] | 3 [2–5] | 3 [1–4] | 2 [1–4] | 2 [1–4] | <.001 | 0
Charlson Comorbidity Index |  |  |  |  | <.001 | 0
1–3 | 23 786 (68) | 261 (55) | 1113 (53) | 7232 (65) | 10 238 (70) | 4942 (73) |  | 
4–7 | 9985 (28) | 168 (36) | 826 (39) | 3441 (31) | 3971 (27) | 1579 (23) |  | 
≥8 | 1444 (4) | 42 (9) | 154 (7) | 532 (5) | 509 (3) | 207 (3) |  | 
Treatments |  |  |  |  |  |  |  | 
ACEi/ARB/ARNi | 32 619 (93) |  | 966 (46) | 10 376 (93) | 14 549 (99) | 6728 (100) | <.001 | 0
ACEi | 16 406 (47) |  | 520 (25) | 5901 (53) | 7423 (50) | 2562 (38) | <.001 | 0
ARB | 11 766 (33) |  | 396 (19) | 4051 (36) | 5263 (36) | 2056 (31) | <.001 | 0
ARNi | 9332 (27) |  | 130 (6) | 1468 (13) | 4315 (29) | 3419 (51) | <.001 | 0
Beta-blocker | 32 413 (92) |  | 969 (46) | 10 273 (92) | 14 443 (98) | 6728 (100) | <.001 | 0
MRA | 21 109 (60) |  | 132 (6) | 1445 (13) | 12 804 (87) | 6728 (100) | <.001 | 0
SGLT2i | 9428 (27) |  | 26 (1) | 316 (3) | 2358 (16) | 6728 (100) | <.001 | 0
Ivabradine | 240 (1) | <10 | 17 (1) | 60 (1) | 110 (1) | 47 (1) | .113 | 0
Loop diuretics | 21 596 (61) | 149 (32) | 1260 (60) | 7023 (63) | 9320 (63) | 3844 (57) | <.001 | 0
Digoxin | 3662 (10) | 22 (5) | 166 (8) | 1154 (10) | 1684 (11) | 636 (9) | <.001 | 0
Nitrate | 6725 (19) | 41 (9) | 385 (18) | 2313 (21) | 2868 (19) | 1118 (17) | <.001 | 0
Number of other CV and non-CV medicationsa |  |  |  | <.001 | 0
0–1 | 2159 (6) | 166 (35) | 140 (7) | 651 (6) | 793 (5) | 409 (6) |  | 
2–4 | 12 139 (34) | 149 (32) | 590 (28) | 3675 (33) | 5147 (35) | 2578 (38) |  | 
5–6 | 8773 (25) | 58 (12) | 500 (24) | 2796 (25) | 3747 (25) | 1672 (25) |  | 
7–8 | 6221 (18) | 40 (8) | 390 (19) | 2058 (18) | 2592 (18) | 1141 (17) |  | 
≥9 | 5923 (17) | 58 (12) | 473 (23) | 2025 (18) | 2439 (17) | 928 (14) |  | 
Device therapya |  |  |  |  |  | <.001 | 0
No | 29 319 (84) | 396 (86) | 1840 (89) | 9768 (88) | 11 794 (80) | 5521 (82) |  | 
CRT/ICD | 5747 (16) | 65 (14) | 235 (11) | 1368 (12) | 2875 (20) | 1204 (18) |  | 

An explorative analysis was performed on the following outcomes: the composite of first HHF and cardiovascular death, first HHF, and cardiovascular death. Survival functions were depicted by cumulative incidence curves. Associations with outcomes of number of prescribed GDMT components, adherence (PDC ≥ vs <80%), and 1-year persistence (yes vs no) were assessed by crude and multivariable Cox regression models, with multivariable models adjusted for covariates reported in Table 1. Results were reported as hazard ratios (HR) with 95% CI. To mitigate potential temporal bias favouring quadruple therapy due to recent prognostic improvements in HFrEF and improved management of comorbidities,16 a sensitivity analysis restricted to patients enrolled 2021–23 was performed for the association between number of prescribed GDMT components and outcomes.

Missing data were handled by multiple imputation (10 sets, R-package _mice_17), including variables labelled in Table 1 in the imputation model. Estimates and standard errors were pooled according to Rubin’s rules. No adjustments were performed for multiple comparisons due to the exploratory nature of the analysis. Two-sided P < .05 were considered statistically significant. The statistical software R 4.5.0 was used for all analyses. The R code for can be found https://github.com/KIHeartFailure/swedehf-hftreat.