Hyperkalemia is a concern in heart failure (HF), especially in older patients with co-morbidities. Previous studies addressing this issue have focused mainly on younger patients. This study was aimed at determining the frequency and predictors of hyperkalemia in older patients with HF undergoing intense medical therapy. Frequency and predictors of hyperkalemia were defined in patients (n = 566) participating in the Trial of Intensified versus Standard Medical Therapy in Elderly Patients with Congestive Heart Failure, in which patients ≥60 years of age were randomized to a standard versus an intensified N-terminal brain natriuretic peptide-guided HF therapy. During an 18-month follow-up 76 patients (13.4%) had hyperkalemia (≥5.5 mmol/L) and 28 (4.9%) had severe hyperkalemia (≥6.0 mmol/L). Higher baseline serum potassium (odds ratio [OR] 2.92 per mmol/L), baseline creatinine (OR 1.11 per 10 μmol/L), gout (OR 2.56), New York Heart Association (NYHA) class (compared to NYHA class II, IV OR 3.08), higher dosage of spironolactone at baseline (OR 1.20 per 12.5 mg/day), and higher dose changes of spironolactone (compared to no dose change: 12.5 mg, OR 1.45; 25 mg, OR 2.52; >25 mg, OR 3.24) were independent predictors for development of hyperkalemia (p <0.05 for all comparisons). In conclusion, hyperkalemia is common in patients ≥60 years of age with HF undergoing intense medical therapy. Risk is increased in patients treated with spironolactone, in addition to patient-specific risk factors such as chronic kidney disease, higher serum potassium, advanced NYHA class, and gout. Careful surveillance of serum potassium and cautious use of spironolactone in patients at risk may help to decrease the incidence of potentially hazardous complications caused by hyperkalemia.
Several studies have described the incidence and risk factors of hyperkalemia in patients with heart failure (HF). However, these studies typically have focused on younger patients with a lower prevalence of relevant co-morbidities known to interfere with potassium homeostasis compared to older patients who represent most patients with HF. Therefore, we aimed to describe the frequency and predictors of hyperkalemia in patients ≥60 years of age with HF taking part in the Trial of Intensified versus Standard Medical Therapy in Elderly Patients with Congestive Heart Failure (TIME-CHF).
Methods
Patients participating in the TIME-CHF were considered for this analysis. Details of the study have been described previously. In brief, patients ≥60 years of age with New York Heart Association (NYHA) class ≥II, a history of HF hospitalization within the previous year, and an N-terminal brain natriuretic peptide (NT-BNP) level >2 times the upper limit of normal were included in the study. Patients with dyspnea not mainly caused by HF, valvular heart disease needing surgery, acute coronary syndrome within 10 days, angina pectoris documented as ischemia by Canadian Cardiovascular Society class ≥II, percutaneous coronary intervention within the previous month, coronary artery bypass graft surgery within the previous 3 months, body mass index ≥35 kg/m 2 , life expectancy <3 years for diseases other than cardiovascular, and serum creatinine level ≥220 μmol/L were excluded. The ethics committee of each center approved the study and each patient gave written informed consent before entering the study (trial registration at: http://isrctn.org , identifier ISRCTN43596477 ).
The TIME-CHF was a prospective randomized controlled multicenter trial addressing the therapeutic management of patients ≥60 years of age with HF. A standard symptom-guided therapy and an intensified NT-BNP–guided medical therapy were compared. After baseline assessment patients were treated according to the study protocol and followed for 18 months with visits after 1 month and 3, 6, 12, and 18 months. Treatments were adjusted at all but the last visit with the attempt to achieve therapeutic goals by the 6-month visit followed by 12 months of outcome observation. Details about therapy uptitration as recommended by the study protocol have been described previously. Detailed clinical and laboratory assessments including measurement of serum potassium concentration were performed per protocol at each follow-up visit. Noncardiac co-morbidities were identified based on medical and hospital case histories. Hyperkalemia and severe hyperkalemia were defined as serum potassium concentrations ≥5.5 and ≥6.0 mmol/L, respectively. Estimated glomerular filtration rate was calculated according to the simplified Modified Diet in Renal Disease study equation. Of the 622 patients included in TIME-CHF, 566 patients who had ≥2 study visits were included in this analysis. Of the remaining 56 patients, 18 died (none owing to known hyperkalemia; 9 because of circulatory failure, 5 because of sudden cardiac death, 2 from a noncardiovascular cause, and 2 from an unknown cause), 37 withdraw consent, and 1 was excluded by investigator decision.
Dosages of angiotensin-converting enzyme (ACE) inhibitors, angiotensin receptor blockers (ARBs), and β blockers at baseline and follow-up were expressed as percent recommended target dosages (e.g., for a target dosage of ramipril 10 mg/day, 5 mg/day represents 50% of the target dose). For the combination of ACE inhibitors and ARBs, the relative dosages were added and expressed as a combined dosage (e.g., 75% of target dosage of ACE inhibitors and 50% of target dosage of ARBs would produce a combined dosage of 125%). Target dosages of the used ACE inhibitors, ARBs, and β blockers as recommended by the guidelines are listed in Table 1 . Although not based on large trials, similar target doses were used for patients with preserved systolic left ventricular function. Dosages of loop diuretics are expressed as equivalent dosages of oral furosemide, where torsemide 10 mg was defined as furosemide 40 mg. This recalculation of dosages was necessary because different ACE inhibitors, ARBs, β blockers, and loop diuretics were used at the discretion of participating investigators. In contrast, spironolactone was the only mineralocorticoid receptor antagonist used in this study and the dosage, therefore, is expressed as milligrams per day.
| Drug | Target Maintenance Dose |
|---|---|
| Angiotensin-converting enzyme inhibitors | |
| Benazepril | 10 mg 2 times/day |
| Captopril | 50 mg 3 times/day |
| Enalapril | 10 mg 2 times/day |
| Lisinopril | 20 mg 1 time/day |
| Quinapril | 10 mg 2 times/day |
| Perindopril | 4 mg 1 time/day |
| Ramipril | 5 mg 2 times/day |
| Cilazapril | 2.5 mg 1 time/day |
| Fosinopril | 20 mg 1 time/day |
| Trandolapril | 4 mg 1 time/day |
| Angiotensin receptor blockers | |
| Losartan | 100 mg 1 time/day |
| Valsartan | 160 mg 1 time/day |
| Irbesartan | 150 mg 1 time/day |
| Candesartan cilexetil | 16 mg 1 time/day |
| Telmisartan | 40 mg 1 time/day |
| Eprosartan | 800 mg 1 time/day |
| β Blockers | |
| Bisoprolol | 10 mg 1 time/day |
| Metoprolol succinate | 200 mg 1 time/day |
| Carvedilol | 25 mg 2 times/day |
| Nebivolol | 10 mg 1 time/day |
| Aldosterone antagonist | |
| Spironolactone | 25–50 mg 1 time/day |
Information on drug treatment and doses was available for every single day throughout the study. Because changes in serum potassium levels were more relevant during the early course of the study, when most therapy adjustments were carried out, we performed a specific analysis focused on the impact of drug uptitration during the first 6 months on changes in serum potassium levels. For the present analysis, maximal dose, maximal dose increase, and mean dose for each class of drugs (i.e., renin–angiotensin–aldosterone inhibitors, β blocker, loop diuretic, spironolactone) were calculated and tested.
Continuous data are presented as mean ± SD or median (interquartile range) as appropriate. Categorical data are presented as number (percentage). The univariate association between the tested variables and incident hyperkalemia was tested by t test for continuous variables with normal distribution, Mann–Whitney U test for ordinal variables or continuous variables without normal distribution, and by Fisher’s test for nominal variables. Independent predictors of hyperkalemia were identified by including variables with a p value <0.1 in univariable analysis in a multivariable logistic regression using forward and backward procedures, which provided similar results. A p value <0.05 was considered statistically significant (2-sided). Analyses were performed using SPSS 15.0 (SPSS, Inc., Chicago, Illinois).
Results
Baseline characteristic of patients included in this study are listed in Table 2 . Notably, the median age was >75 years, and the prevalence of co-morbidities potentially interfering with potassium homeostasis was frequent, with >1 of 2 patients having a history of chronic renal failure and >1 of 3 having diabetes mellitus.
| Variable | Hyperkalemia | p Value | |
|---|---|---|---|
| No | Yes | ||
| (n = 490) | (n = 76) | ||
| Age (years) | 76.4 ± 7.6 | 78.0 ± 7.2 | 0.08 |
| Men | 285 (58%) | 53 (70%) | 0.06 |
| Randomization to N-terminal brain natriuretic peptide-guided therapy | 252 (51%) | 33 (43%) | 0.22 |
| Body mass index (kg/m 2 ) | 25.6 ± 4.3 | 25.8 ± 5.1 | 0.73 |
| Systolic blood pressure (mm Hg) | 122 ± 20 | 120 ± 17 | 0.36 |
| New York Heart Association Classes II/III/IV | 129/306/55 (26%/62%/11%) | 15/44/17 (20%/58%/22%) | 0.02 |
| Left ventricular ejection fraction (%) | 35 ± 13 | 34 ± 14 | 0.52 |
| Heart failure with preserved left ventricular ejection fraction | 90 (18%) | 14 (18%) | 1.00 |
| Coronary artery disease | 250 (51%) | 44 (58%) | 0.27 |
| Atrial fibrillation | 158 (33%) | 27 (36%) | 0.60 |
| Hypertension | 365 (75%) | 55 (72%) | 0.67 |
| Diabetes mellitus | 158 (32%) | 35 (46%) | 0.03 |
| Hypercholesterolemia | 237 (48%) | 43 (57%) | 0.22 |
| Peripheral occlusive artery disease | 93 (19%) | 18 (24%) | 0.35 |
| Cerebrovascular disease | 76 (16%) | 14 (18%) | 0.50 |
| Chronic obstructive pulmonary disease | 91 (19%) | 19 (25%) | 0.21 |
| Kidney disease | 257 (52%) | 57 (75%) | <0.001 |
| Anemia | 126 (26%) | 29 (38%) | 0.03 |
| Gout | 36 (7%) | 13 (17%) | 0.01 |
| Cancer | 71 (15%) | 8 (11%) | 0.48 |
| Osteoarthritis | 138 (28%) | 21 (28%) | 1.00 |
| N-terminal brain natriuretic peptide (pg/ml) | 3,555 (1,829–6,520) | 4,854 (2,367–7,467) | 0.02 |
| Creatinine (mmol/L) | 112 ± 36 | 135 ± 41 | <0.001 |
| Estimated glomerular filtration rate (ml/min/1.73 m 2 ) | 55 ± 19 | 46 ± 18 | <0.001 |
| Urea (mmol/L) | 10.8 ± 5.0 | 13.9 ± 6.0 | <0.001 |
| Potassium (mmol/L) | 4.1 ± 0.5 | 4.4 ± 0.6 | <0.001 |
| Sodium (mmol/L) | 140 ± 8 | 139 ± 4 | 0.83 |
| Hemoglobin (mmol/L) | 131 ± 18 | 128 ± 17 | 0.10 |
| β Blockers | 383 (78%) | 57 (75%) | 0.55 |
| Angiotensin-converting enzyme inhibitors or angiotensin receptor blockers | 457 (93%) | 70 (92%) | 0.63 |
| Spironolactone | 176 (36%) | 37 (49%) | 0.04 |
| Loop diuretics | 451 (92%) | 72 (95%) | 0.49 |
| Dose of β blockers, median (interquartile range of target dosage) | 25 (6.25–50) | 25 (1.25–50) | 0.30 |
| Dose of angiotensin-converting enzyme inhibitors or angiotensin receptor blockers, median (interquartile range of target dosage) | 50 (25–67) | 50 (25–67) | 0.97 |
| Dose of spironolactone, median (interquartile range of target dosage) | 0 (0–25) | 0 (0–25) | 0.01 |
| Dose of loop diuretics, median (interquartile range of furosemide equivalents) | 60 (40–80) | 80 (40–120) | 0.02 |
During the entire study period, 76 patients (13.4%) had hyperkalemia (≥5.5 mmol/L) and 28 (4.9%) had severe hyperkalemia (≥6.0 mmol/L). Figure 1 illustrates the frequency of hyperkalemia during the 18-month course of the study, showing some increase in hyperkalemia early after intensifying therapy and late during the study course.
Results of univariable analysis to identify predictors at baseline for development of hyperkalemia are presented in Table 2 . Patients developing hyperkalemia were more likely to be in NYHA class III or IV and to have a history of diabetes mellitus, gout, and renal failure, higher serum concentrations of creatinine, urea, potassium, and NT-BNP, and a lower estimated glomerular filtration rate. At baseline dosages of spironolactone and loop diuretics were higher in patients developing hyperkalemia than in those who did not. In the 76 patients developing hyperkalemia, the average estimated glomerular filtration rate at the time of hyperkalemia was significantly lower compared to visits when the same patients did not have hyperkalemia (34 ± 12 vs 43 ± 15 ml/min/1.73 m 2 , p <0.001). In a multivariable logistic regression model, serum potassium and creatinine, history of gout, NYHA class, and dose of spironolactone at baseline were independent predictors of hyperkalemia ( Table 3 ). Figure 2 depicts frequencies of hyperkalemia according to presence of renal failure, serum potassium levels, NYHA class, history of gout, and dosage of spironolactone prescribed at baseline.
