Serum osmolality may fluctuate with neurohormonal activation and in response to certain therapeutics in patients with heart failure (HF). The clinical relevance of osmolality in patients with HF has not been defined. In this post hoc analysis of the Efficacy of Vasopressin Antagonism in Heart Failure: Outcome Study with Tolvaptan trial, we analyzed serum osmolality measured at discharge in 3,744 patients hospitalized for HF and reduced ejection fraction (EF ≤40%). Median follow-up was 9.9 months. The association between discharge osmolality and all-cause mortality (ACM) and composite cardiovascular mortality or HF hospitalization was nonlinear; and thus, patients were divided into low (≤284), normal (285 to 300), and high (≥300 mOsm/kg) osmolality. Median serum osmolality at discharge was 297 (290 to 304) mOsm/kg. Patients in the low osmolality group (n = 454,12.1%) were more likely to be younger, men, have lower rates of hypertension, coronary artery disease, chronic kidney disease, diabetes, and have lower serum sodium, creatinine, systolic blood pressure, and EF (all p <0.001). Low discharge osmolality was associated with higher ACM (low 29.3%; normal 23.6%; high 25.2%; p = 0.04) and the composite endpoint (low 45.6%; normal 39.3%; high 41.8%; p = 0.04). After risk adjustment, a 15 mOsm/kg decrease in osmolality was predictive of higher ACM (hazard ratio 1.61, 95% CI 1.19 to 2.17) and the composite endpoint (hazard ratio 1.37, 95% CI 1.06 to 1.75) in the low osmolality group. These associations were not seen in patients with normal or high osmolality. Interaction analyses for tolvaptan treatment were nonsignificant (p >0.4). In conclusion, low discharge serum osmolality was independently predictive of worse postdischarge mortality and readmission. Further study is required to clarify the clinical utility of serum osmolality in hospitalized patients with HF.
Patients hospitalized for worsening chronic heart failure (HF) face high rates of postdischarge adverse outcomes, with mortality exceeding 25% within 12 months of discharge. The identification of modifiable targets of therapy in hospitalized patients with HF is an unmet clinical need. We recently demonstrated that serum osmolality can be accurately determined by parameters included in routine laboratory tests using the formula: [2 × serum sodium (mEq/L) + blood urea nitrogen (BUN, mg/dl)/2.8 + glucose (mg/dl)/18]. Serum osmolality appears to increase during and after hospitalization for HF, driven largely by changes in serum sodium and BUN and responds robustly to vasopressin antagonism with tolvaptan therapy. However, the clinical profile and prognostic value of discharge serum osmolality have not been defined in HF. In this post hoc analysis of the EVEREST (Efficacy of Vasopressin Antagonism in Heart Failure: Outcome Study with Tolvaptan) trial, we aimed to: (1) define the clinical phenotype and predictive value of discharge osmolality and (2) determine whether this risk is modifiable with tolvaptan.
Methods
The study design and primary results of the EVEREST trial have been reported previously. Briefly, EVEREST was an international, prospective, randomized, double-blind, placebo-controlled trial designed to examine the short- and long-term efficacy and safety of tolvaptan, a selective vasopressin-2 receptor antagonist. Patients aged ≥18 years hospitalized for worsening HF with New York Heart Association class III or IV symptoms, reduced ejection fraction (EF) ≤40%, and presenting with 2 or more signs or symptoms of volume overload were eligible for enrollment. Relevant exclusion criteria included supine systolic blood pressure <90 mm Hg, serum creatinine >3.5 mg/dl, requirement for hemofiltration or dialysis, refractory, end-stage HF, or a life expectancy of <6 months.
Patients were randomized within 48 hours of admission to receive either 30 mg of oral tolvaptan once daily or matching placebo, both in addition to standard therapy, for a minimum of 60 days. Background HF therapy was left to the discretion of the treating physician. The EVEREST trial was conducted in accordance with the Declaration of Helsinki, the protocol was independently approved by the institutional review board or ethics committee at each participating center, and written informed consent was obtained from all participants.
Osmolality, expressed in mOsm/kg, was measured at the time of enrollment (referred to as “baseline” up to 48 hours after admission), during hospitalization, and at discharge. Serum osmolality was directly measured using freezing point depression methods and cross-validated across 5 central core laboratories. The primary predictor variable, discharge serum osmolality, was defined as the serum osmolality value on in-hospital day 7 or discharge, whichever occurred first. For patients with missing discharge osmolality measurements, the latest in-hospital osmolality measurement was used.
The association between discharge osmolality and clinical endpoints was nonlinear; and therefore, it was determined separately for the 3 groups: “low” (≤285 mOsm/kg), “normal” (285 to 300 mOsm/kg), and “high” (≥300 mOsm/kg). Cutoffs were determined a priori using standard core laboratory values for abnormal parameters and identified approximately 50% of patients as having “normal” discharge osmolality. In-hospital change was defined as discharge/day 7 osmolality minus the baseline measurement (with a positive value reflecting a relative increase in osmolality, whereas a negative value refers to a relative decrease).
An independent and blinded committee adjudicated the prespecified coprimary endpoints, all-cause mortality (ACM) and the composite of cardiovascular (CV) mortality and HF hospitalization. For descriptive purposes, both tolvaptan and placebo groups were combined for the present analysis to define the clinical profiles of low, normal, and high discharge osmolality measurements. The value of analyzing the predictive value of discharge osmolality separately for each treatment group (tolvaptan and placebo) was determined using interaction analyses.
Baseline and discharge characteristics and the occurrence of clinical events were compared across discharge osmolality groups using the analysis of variance, Kruskal–Wallis, and chi-square tests, as appropriate. Kaplan–Meier curves by discharge osmolality group were constructed for both primary endpoints and compared using log-rank tests. Effect sizes associated with serum osmolality as a continuous variable were also assessed using Cox proportional hazards regression models and reported as hazard ratios (HRs) with 95% confidence intervals (CI).
All multivariate models were adjusted for 25 preselected discharge variables: age, gender, medical comorbidities, ischemic HF origin, a history of previous HF hospitalization, New York Heart Association class IV symptoms, EF, QRS duration and atrial fibrillation/flutter on electrocardiogram, discharge systolic blood pressure, discharge laboratory findings (natriuretic peptides, serum sodium, BUN, and blood glucose), and discharge medication use. Fully conditional specification was used for imputation of missing covariate data (∼30% for natriuretic peptides, <5% for QRS duration, albumin, and renal parameters, and ≤2% for all other variables). All analyses were performed using SAS, version 9.2 (Cary, North Carolina).
Results
The EVEREST program enrolled 4,133 patients, of which 3,744 (90.6%) had discharge osmolality data available. Patients with missing osmolality data were excluded (n = 389). Median baseline osmolality was 295 (interquartile range [IQR] 288 to 301) mOsm/kg and discharge/day 7 osmolality was 297 (IQR 290 to 304) mOsm/kg, with a median in-hospital change of 3 (IQR −4 to 10) mOsm/kg. Patients were divided based on discharge osmolality into low (≤284 mOsm/kg; n = 454; 12.1%), normal (285 to 300 mOsm/kg; n = 1,862; 49.7%), and high (≥300 mOsm/kg; n = 1,428; 38.1%) groups. The low discharge osmolality group experienced a median in-hospital decrease in osmolality by 8 mOsm/kg (IQR −16 to 0), whereas the high discharge osmolality group experienced a median in-hospital increase by 8 mOsm/kg (IQR 2 to 15; p <0.001).
Increasing discharge osmolality was associated with higher percentages of tolvaptan treatment assignment (27.5% vs 46.2% vs 61.9%; p <0.001; Table 1 ). Serum sodium, BUN, glucose levels, and body weight were lowest in the low osmolality group at enrollment and discharge (all p <0.001). In-hospital change in weight was similar across groups (−2.3 kg in low osmolality; −2.5 kg in normal osmolality; −2.7 kg in high osmolality; p = 0.31). Patients with the lowest osmolality were younger and more likely to be men, nonwhite, and enrolled from South America (all p ≤0.002; Table 1 ). Lower osmolality was associated lower systolic blood pressures and higher heart rates (p <0.001). Natriuretic peptides were elevated in all patients but appear to follow a U-shaped relationship with osmolality. Patients with the lowest osmolality had lower EF (p <0.001) than the other groups. These trends in clinical and laboratory parameters largely persisted at discharge. Patients with lower osmolality had a lower prevalence of ischemic HF origin, coronary artery disease, and other cardiac comorbidities (p ≤0.001; Table 1 ). Patient with low osmolality had lower rates of discharge prescription of β blockers, calcium channel blockers, nitrates, and statins (all p ≤0.007) but had higher rates of angiotensin-converting enzyme inhibitor/angiotensin II receptor blocker (p = 0.006) and mineralocorticoid receptor antagonist prescription (p <0.001) compared with higher osmolality groups ( Table 2 ).
| Variable | Serum Osmolality at Discharge or Day 7 (mOsm/kg) | |||
|---|---|---|---|---|
| ≤284 (n=454) | 285-300 (n=1862) | ≥300 (n=1428) | p | |
| Tolvaptan assignment | 125 (27.5%) | 860 (46.2%) | 884 (61.9%) | <0.001 |
| Demographic characteristics | ||||
| Age (years), mean±SD | 60.9±14.1 | 64.6±11.5 | 68.5±10.3 | <0.001 |
| Male | 357 (78.6%) | 1419 (76.2%) | 1025 (71.8%) | 0.002 |
| Race | <0.001 | |||
| Non-Hispanic White | 354 (78%) | 1589 (85.3%) | 1277 (89.4%) | |
| Black | 39 (8.6%) | 139 (7.5%) | 87 (6.1%) | |
| Hispanic | 44 (9.7%) | 92 (4.9%) | 46 (3.2%) | |
| Other | 17 (3.7%) | 42 (2.3%) | 17 (1.2%) | |
| Region | <0.001 | |||
| Eastern Europe | 124 (27.3%) | 770 (41.4%) | 633 (44.3%) | |
| North America | 111 (24.4%) | 577 (31%) | 418 (29.3%) | |
| South America | 182 (40.1%) | 323 (17.3%) | 138 (9.7%) | |
| Western Europe | 37 (8.1%) | 192 (10.3%) | 239 (16.7%) | |
| Enrollment Signs, Symptoms, and Laboratory Findings | ||||
| Sodium (mEq/L), median (IQR) | 138 (135-141) | 139 (137-142) | 141 (138-143) | <0.001 |
| Blood urea nitrogen (mg/dL), median (IQR) | 23 (18-30) | 24 (18-31) | 30 (22-42) | <0.001 |
| Glucose (mg/dL), median (IQR) | 115 (97-144) | 119 (98-157) | 128 (104-170) | <0.001 |
| Weight (kg), mean±SD | 80±18.4 | 83.7±19.3 | 84.2±18.1 | <0.001 |
| Systolic blood pressure (mmHg), mean±SD | 115.4±17.7 | 119.8±19.3 | 123±20 | <0.001 |
| Diastolic blood pressure (mmHg), mean±SD | 71.7±11.8 | 72.7±12.6 | 73±12.9 | 0.185 |
| Heart rate (bpm), mean±SD | 81.6±16.2 | 80.1±15.5 | 78.7±15.3 | 0.001 |
| Jugular venous distension≥10cm | 153 (34.5%) | 494 (27.3%) | 332 (23.8%) | <0.001 |
| Rales | 359 (80.1%) | 1498 (82.1%) | 1149 (81.6%) | 0.632 |
| Peripheral edema ∗ | 365 (81.5%) | 1451 (79.4%) | 1142 (81.1%) | 0.393 |
| Dyspnea | 400 (90.1%) | 1665 (91.2%) | 1287 (91.5%) | 0.667 |
| Ejection fraction (%), mean±SD | 25.5±8.2 | 27.3±8.1 | 28.5±7.8 | <0.001 |
| Creatinine (mg/dL), median (IQR) | 1.1 (1-1.4) | 1.2 (1-1.4) | 1.4 (1.1-1.8) | <0.001 |
| Estimated glomerular filtration rate (mL/min/1.73 m2), median (IQR) † | 61.7 (49.8-76.5) | 59.2 (45.8-72.8) | 46.4 (33.2-61.2) | <0.001 |
| B-type natriuretic peptide (pg/ml), median (IQR) | 762 (324.1-1688.8) | 655.9 (276-1388.3) | 698 (283-1552) | 0.055 |
| N-terminal-pro B-type natriuretic peptide (pg/ml), median (IQR) | 4819 (2616.6-9984) | 4079.3 (1781-8322) | 5287.4 (2472-10396) | 0.001 |
| Albumin (g/dL), median (IQR) | 3.7 (3.4-4) | 3.8 (3.4-4.1) | 3.8 (3.4-4.1) | 0.011 |
| Discharge Signs, Symptoms, and Laboratory Findings | ||||
| Sodium (mEq/L), median (IQR) | 137 (133-141) | 139 (137-142) | 142 (139-145) | <0.001 |
| Blood urea nitrogen (mg/dL), median (IQR) | 23 (18-31) | 25 (19-33) | 34 (24-49) | <0.001 |
| Glucose (mg/dL), median (IQR) | 113 (94-142) | 119 (99-157) | 139 (108-193) | <0.001 |
| Weight (kg), mean±SD | 77±18.1 | 80.7±18.9 | 80.8±17.6 | <0.001 |
| In-hospital change in weight (kg), median (IQR) | -2.3 (-4.4 – -1.0) | -2.5 (-4.6 – -1) | -2.7 (-4.9 – -1) | 0.306 |
| Systolic blood pressure (mmHg), mean±SD | 110.6±17 | 114.9±16.8 | 118.7±16.9 | <0.001 |
| Diastolic blood pressure (mmHg), mean±SD | 68.2±10.8 | 69.8±11 | 70.4±11 | 0.001 |
| Heart rate (bpm), mean±SD | 77±12.6 | 74.6±12.2 | 73.5±12.1 | <0.001 |
| Jugular venous distension≥10cm | 21 (6.3%) | 62 (4.2%) | 33 (3.1%) | 0.032 |
| Rales | 54 (16.2%) | 299 (20.3%) | 221 (20.7%) | 0.178 |
| Peripheral edema ∗ | 111 (33.2%) | 524 (35.6%) | 394 (36.8%) | 0.481 |
| Dyspnea | 58 (17.4%) | 231 (15.7%) | 172 (16.1%) | 0.758 |
| Creatinine (mg/dL), median (IQR) | 1.1 (1-1.4) | 1.2 (1-1.5) | 1.5 (1.2-1.9) | <0.001 |
| B-type natriuretic peptide (pg/ml), median (IQR) | 488.1 (192-1155) | 431 (176-950) | 435 (187.2-982) | 0.227 |
| N-terminal-pro B-type natriuretic peptide (pg/ml), median (IQR) | 3084.6 (1643-5893) | 2368 (996.7-4927) | 3172 (1473-6837) | <0.001 |
| Albumin (g/dL), median (IQR) | 3.8 (3.4-4.1) | 3.9 (3.5-4.2) | 3.9 (3.5-4.2) | 0.005 |
| Initial Electrocardiogram | ||||
| QRS duration (ms), median (IQR) | 123 (99-154) | 123 (97-149) | 122 (97-151) | 0.701 |
| Atrial Fibrillation | 120 (26.5%) | 557 (29.9%) | 404 (28.3%) | 0.293 |
| Medical history | ||||
| Previous heart failure hospitalization | 337 (74.6%) | 1464 (79%) | 1149 (80.7%) | 0.02 |
| Ischemic heart failure etiology | 248 (56%) | 1163 (63.3%) | 1023 (72%) | <0.001 |
| New York Heart Association class IV | 203 (44.7%) | 707 (38%) | 568 (39.8%) | 0.031 |
| Coronary artery disease | 264 (58.1%) | 1281 (68.8%) | 1096 (76.9%) | <0.001 |
| Previous myocardial infarction | 197 (43.4%) | 900 (48.4%) | 805 (56.4%) | <0.001 |
| Hypertension | 273 (60.1%) | 1291 (69.3%) | 1092 (76.5%) | <0.001 |
| Hypercholesterolemia | 185 (40.8%) | 896 (48.3%) | 726 (51.2%) | 0.001 |
| Diabetes mellitus | 112 (24.7%) | 641 (34.4%) | 680 (47.6%) | <0.001 |
| Chronic kidney disease | 66 (14.5%) | 393 (21.1%) | 533 (37.3%) | <0.001 |
| Peripheral vascular disease | 75 (16.6%) | 368 (19.8%) | 341 (23.9%) | 0.001 |
| Previous coronary artery bypass graft surgery | 88 (19.4%) | 374 (20.1%) | 323 (22.6%) | 0.141 |
| Previous percutaneous coronary intervention | 70 (15.4%) | 334 (17.9%) | 262 (18.3%) | 0.354 |
| Automatic implantable cardioverter-defibrillator | 50 (11%) | 263 (14.1%) | 213 (14.9%) | 0.113 |
| Chronic obstructive pulmonary disease | 42 (9.3%) | 191 (10.3%) | 144 (10.1%) | 0.815 |
∗ Peripheral edema was defined as slight/moderate/marked pedal or sacral edema.
† Glomerular filtration rate estimated by Cockcroft-Gault formula.
| Serum Osmolality at Discharge or Day 7 (mOsm/kg) | ||||
|---|---|---|---|---|
| ≤284 (n=454) | 285-300 (n=1862) | ≥300 (n=1428) | p | |
| Baseline medication use | ||||
| Aspirin | 249 (55%) | 1051 (56.5%) | 807 (56.7%) | 0.807 |
| Angiotensin-converting enzyme inhibitor / Angiotensin II receptor blocker | 400 (88.3%) | 1590 (85.5%) | 1189 (83.5%) | 0.034 |
| Mineralocorticoid receptor antagonist | 303 (66.9%) | 1078 (58%) | 732 (51.4%) | <0.001 |
| Beta-Blockers | 296 (65.3%) | 1294 (69.6%) | 1066 (74.9%) | <0.001 |
| Calcium Channel Blocker | 29 (6.4%) | 197 (10.6%) | 173 (12.1%) | 0.003 |
| Warfarin | 155 (34.2%) | 683 (36.7%) | 494 (34.7%) | 0.386 |
| Digoxin | 245 (54.1%) | 981 (52.7%) | 594 (41.7%) | <0.001 |
| Diuretics | 439 (96.9%) | 1816 (97.6%) | 1388 (97.5%) | 0.676 |
| Nitrates | 142 (31.3%) | 676 (36.3%) | 632 (44.4%) | <0.001 |
| Statin | 127 (28%) | 639 (34.4%) | 520 (36.5%) | 0.004 |
| Discharge medication use | ||||
| Aspirin | 230 (52.2%) | 1007 (54.6%) | 756 (53.8%) | 0.626 |
| Angiotensin-converting enzyme inhibitor / Angiotensin II receptor blocker | 390 (88.4%) | 1587 (86.1%) | 1167 (83%) | 0.006 |
| Mineralocorticoid receptor antagonist | 321 (72.8%) | 1188 (64.5%) | 778 (55.3%) | <0.001 |
| Beta-Blockers | 305 (69.2%) | 1377 (74.7%) | 1105 (78.6%) | <0.001 |
| Calcium Channel Blocker | 23 (5.2%) | 152 (8.2%) | 140 (10%) | 0.007 |
| Warfarin | 165 (37.4%) | 763 (41.4%) | 545 (38.8%) | 0.163 |
| Digoxin | 242 (54.9%) | 944 (51.2%) | 578 (41.1%) | <0.001 |
| Diuretics | 418 (94.8%) | 1723 (93.5%) | 1308 (93%) | 0.428 |
| Nitrates | 105 (23.8%) | 522 (28.3%) | 524 (37.3%) | <0.001 |
| Statin | 124 (28.1%) | 631 (34.2%) | 510 (36.3%) | 0.007 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
