Author + information
- Received January 1, 2019
- Revision received February 14, 2019
- Accepted February 20, 2019
- Published online April 29, 2019.
- Pieter Martens, MDa,b,
- Matthias Dupont, MDa,
- Frederik Hendrik Verbrugge, MD, PhDa,
- Kevin Damman, MD, PhDc,
- Nicolas Degrysea,
- Petra Nijst, MD, PhDa,
- Carmen Reynders, MSSCd,
- Joris Penders, MD PhDd,
- W.H. Wilson Tang, MDe,
- Jeffrey Testani, MDf and
- Wilfried Mullens, MD, PhDa,g,∗ ()
- aDepartment of Cardiology, Ziekenhuis Oost-Limburg, Genk, Belgium
- bDoctoral School for Medicine and Life Sciences, Hasselt University, Diepenbeek, Belgium
- cDepartment of Cardiology, University of Groningen, University Medical Centrum Groningen, the Netherlands
- dDepartment of Laboratory Medicine, Ziekenhuis Oost-Limburg, Genk, Belgium
- eDepartment of Cardiovascular Medicine, Heart and Vascular Institute, Cleveland Clinic, Cleveland, Ohio
- fDepartment of Cardiovascular Medicine, Yale University, New Haven, Connecticut
- gBiomedical Research Institute, Faculty of Medicine and Life Sciences, Hasselt University, Diepenbeek, Belgium
- ↵∗Address for correspondence:
Dr. Wilfried Mullens, Department of Cardiology, Ziekenhuis Oost-Limburg, Schiepse Bos 6, 3600 Genk, Belgium.
Objectives This study sought to determine the relationship between urinary sodium (Una) concentration and the pathophysiologic interaction with the development of acute heart failure (AHF) hospitalization.
Background No data are available on the longitudinal dynamics of Una concentration in patients with chronic heart failure (HF), including its temporal relationship with AHF hospitalization.
Methods Stable, chronic HF patients with either reduced or preserved ejection fraction were prospectively included to undergo prospective collection of morning spot Una samples for 30 consecutive weeks. Linear mixed modeling was used to assess the longitudinal changes in Una concentration. Patients were followed for the development of the clinical endpoint of AHF.
Results A total of 80 chronic HF patients (71 ± 11 years of age; an N-terminal pro–B-type natriuretic peptide [NT-proBNP] concentration of 771 [interquartile range: 221 to 1,906] ng/l; left ventricular ejection fraction [LVEF] 33 ± 7%) prospectively submitted weekly pre-diuretic first void morning Una samples for 30 weeks. A total of 1,970 Una samples were collected, with mean Una concentration of 81.6 ± 41 mmol/l. Sodium excretion remained stable over time on a population level (time effect p = 0.663). However, interindividual differences revealed the presence of high (88 mmol/l Una [n = 39]) and low (73 mmol/l Una [n = 41]) sodium excreters. Only younger age was an independent predictor of high sodium excretion (odds ratio [OR]: 0.91; 95% confidence interval [CI]: 0.83 to 1.00; p = 0.045 per year). During 587 ± 54 days of follow-up, 21 patients were admitted for AHF. Patients who developed AHF had significantly lower Una concentrations (F[1.80] = 24.063; p < 0.001). The discriminating capacity of Una concentration to detect AHF persisted after inclusion of NT-proBNP and estimated glomerular filtration rate (eGFR) measurements as random effects (p = 0.041). Furthermore, Una concentration dropped (Una = 46 ± 16 mmol/l vs. 70 ± 32 mmol/l, respectively; p = 0.003) in the week preceding the hospitalization and returned to the individual’s baseline (Una = 71 ± 22 mmol/l; p = 0.002) following recompensation, while such early longitudinal changes in weight and dyspnea scores were not apparent in the week preceding decompensation.
Conclusions Overall, Una concentration remained relatively stable over time, but large interindividual differences existed in stable, chronic HF patients. Patients who developed AHF exhibited a chronically lower Una concentration and exhibited a further drop in Una concentration during the week preceding hospitalization. Ambulatory Una sample collection is feasible and may offer additional prognostic and therapeutic information.
Chronic sodium retention is a hallmark of the heart failure (HF) syndrome, leading to extracellular volume expansion, which contributes to the occurrence of congestion and clinical instability (1). Although all ingested sodium is completely absorbed in the gastrointestinal tract, a neutral sodium balance is retained in a compensated patient through tightly regulated renal sodium excretion and interstitial buffering (2,3). However, in HF, the renal natriuretic response to any given volume status is hampered by unrestrained neurohumoral activation and hemodynamic alterations (4,5). An increasing body of evidence has investigated the role of urinary sodium (Una) concentration (in spot samples and continuous urine collections) during episodes of acute HF (AHF), linking low Una concentration to a diminished diuretic response, ongoing congestion, and an increased risk for developing HF readmission or cardiovascular mortality (6–11). However, no (longitudinal) data are available for Una concentration in patients with chronic HF. Additionally, a relationship between alterations in chronic renal sodium excretion and the risk of developing AHF has not been established in chronic HF. Therefore, insights into chronic renal sodium handling may help to understand the instigating mechanisms of stable HF patients who develop AHF. This study consisted of performing an exploratory and hypothesis-generating prospective analysis to assess profiles of Una concentration in patients with stable HF. Furthermore, the relationship between ambulatory spot Una concentration and the risk of developing AHF resulting in hospitalization was assessed.
Study population recruitment
Patients were prospectively enrolled in a single tertiary HF clinic between January 2016 and October 2016. Subjects were eligible if they were older than 18 years of age and able to provide written informed consent. Inclusion criteria consisted of: 1) stable HF with no HF hospitalization in the previous 3 months; 2) a diagnosis of HF with reduced ejection fraction (HFrEF) defined as previously symptomatic HF with a left ventricular ejection fraction (LVEF) below 45% or a diagnosis of HF with preserved ejection fraction (HFpEF) defined as previously symptomatic HF with a LVEF above 50% and, additionally, an N-terminal pro–B-type natriuretic peptide (NT-proBNP) concentration of >1,000 ng/l and echocardiographic signs of diastolic dysfunction (increased left ventricular wall mass or diastolic dysfunction >grade I or an increased left atrium volume index with cutoff values as defined by European Society of Cardiology (ESC) HF guidelines (12); and 3) receiving stable doses of guideline-recommended disease-modifying therapies for at least 3 months, in the case of HFrEF. Exclusion criteria consisted of: 1) the clinical impression by the study team that the patient was unable to adhere to the study protocol; 2) patients were undergoing or planned for renal replacement therapy; and 3) there was an absence of a freezer at home. Patients were followed by a multidisciplinary HF team in close collaboration with primary care physicians and were instructed about a low-sodium diet according to ESC HF guidelines (12). The study was carried out in accordance with tenets of the Declaration of Helsinki. The study protocol was approved by the local institutional review board, and all patients provided written informed consent. The manuscript was drafted according to STROBE (STrengthening the Reporting of OBservational studies in Epidemiology) guidelines for observational studies (13).
Baseline data collection
After patients completed their informed consent, they underwent standardized baseline evaluation consisting of a history of functional status, a medical history and baseline medication evaluation, a physical examination, an electrocardiogram, a transthoracic echocardiogram, blood sampling, and 24-h urine collection.
Weekly urine sampling and storage
During the baseline visit, patients received a thorough explanation of how to collect ambulatory spot urine samples. Briefly, patients were instructed to collect a once-weekly first void morning urine sample. The patients who were taking diuretic medicines in the morning were instructed to take the diuretic only after collection of the first morning void. Patients were instructed to always collect the morning urine sample on the similar day of consecutive weeks (only weekdays were allowed, not Saturday or Sunday). Urine was collected in a disposable urine collection cup. Afterward, a spot sample was aspirated from the sealed collection cup by using the aspiration port (Online Figure 1). The patients were to immediately place the labeled vacuum tube containing the urine in their freezer (with a commercial standard of approximately −18°C). On the same morning, patients filled in a biometric data questionnaire consisting of dyspnea scoring (visual analog scale [VAS]), weight, blood pressure, and loop diuretic dose and indicated if obvious changes in diet compared with the previous week were made. Both the urinary tube and the biometric data questionnaire contained an identification corresponding to the correct week (e.g., week 1, week 2, and so forth). Just before freezing the urine sample, the patient signed a safety check confirming that both the urine sample and the biometric questionnaire had been double-checked with the appropriate week’s labeling.
Clinical follow-up and sample analysis
Patients returned for ambulatory follow-up at pre-defined intervals, namely, weeks 12, 24, and 30. Week 30 was the final date of ambulatory follow-up, but patients were followed prospectively for the development of AHF or all-cause mortality. A visual synopsis of the study is shown in Figure 1. At follow-up, patients brought the frozen urine samples to the clinic in a commercially available freezing bag. Upon arrival at the clinic, urine samples were immediately stored at −18°C, until the day of analysis. Urine samples were assessed in batches for Una concentration (Cobas-8000 analyzer, Roche Diagnostics, Basel, Switzerland). Additionally, during the follow-up visit, patients underwent physical examination, laboratory assessments, and a repeat 24-h urine collection. All patients’ electronic health records were tagged to prospectively assess development of AHF hospitalizations and all-cause mortality. An AHF hospitalization was defined as a hospitalization lasting at least 24 h with either 2 signs or symptoms of congestion necessitating the use of intravenous diuresis or, as signs of hypoperfusion, necessitating the use of intravenous vasoactive drugs. Censoring of the prospective events occurred on February 1, 2018.
Relationship between spot and 24-h urine collection sodium
In line with published studies of AHF, spot Una concentration (expressed in mmol/l) was the primary marker of interest for the current analysis (6–11), especially because spot samples are more convenient to obtain than 24-h urine collection. However, to ensure that spot samples conveyed similar information as 24-h urine collection in the patient populations of interest (patients who did or did not develop AHF), the relationship between spot Una concentration and continuous collection of Una concentration was compared in these patients. The rationale for doing so was based on the observation that sicker patients, who were more likely to develop an AHF episode, could exhibit enhanced nocturnal diuresis, which might have diluted the morning spot Una concentration in comparison to the 24-h Una concentration sample.
Continuous variables are mean ± SD if normally distributed or median (interquartile range [IQR]) if not normally distributed. Normality was checked by using the Shapiro-Wilk test, Q-Q plots, skewness, and kurtosis. Categorical data were expressed as numbers and percentages and compared by using the Pearson chi-square test or Fisher’s exact test when appropriate. Continuous baseline variables were compared using Student’s t-test or Mann-Whitney U test as appropriate. Longitudinal Una profiles were assessed using linear mixed modeling for repeated measures. Because missing values were expected to occur during follow-up in the repeated-measure design, mixed modeling was preferred to repeated-measures analysis of variance. The time effect was modeled using first-order autoregressive. This method was chosen because it could be assumed that measurements of sodium taken closely together are more correlated to each other. Models were built to investigate the fixed effect of time, group (development of AHF hospitalization), and a [group*time] effect. Additionally, random effects of intercept and [HF*time] were included in the model. The time variable was assessed in both a linear and a quadric fashion to assess the appropriateness of a linear mixed models versus a nonlinear mixed model. Fixed effects were analyzed using a sum-of-squares type III. Post hoc testing of the random effects was performed using a t-test to identify different profiles in slope and intercept of the curve (division between high and low sodium excreters). Additionally, to determine if Una concentration was predictive, in addition to NT-proBNP and eGFR, these variables were included as random effects in additional analysis. For patients experienced an AHF hospitalization during the period of sodium sampling, the Una concentrations before decompensation and the week of decompensation or week after decompensation were analyzed by using a paired t-test. To assess whether spot Una concentrations in patients who developed an AHF hospitalization conveyed information similar to that of continuously collected Una samples, patients with and without an AHF episode were compared. The spot Una concentrations from all individually available samples were averaged and correlated with the Una concentration from the 3 consecutive 24-h urine collections by using Pearson’s correlation. The strength of the correlations were compared following R-to-Z Fisher transformation. Statistical significance was always set at a 2-tailed probability level of <0.05. Statistics were performed using SPSS version 22 software (IBM, Armonk, New York).
A total of 100 patients were prospectively included between January 2016 and October 2016. However, 10 patients who underwent baseline evaluation and signed an informed consent did not return for study follow-up. Additionally, 10 patients were excluded because the ambulatory collection of urinary samples were performed incorrectly (missing week labels) or inconsistently (multiple samples missing). Therefore, the final patient population consisted of 80 patients. Baseline characteristics are reflected in the left column of Table 1. Patients were predominantly male, and most patients had HFrEF and an ischemic etiology. Patients had moderately impaired renal function (eGFR of 54.7 ± 19.4 ml/min per 1.73 m2) and were only mildly symptomatic at baseline, as illustrated by the large proportion of patients functioning in NYHA functional class II and NT-proBNP concentration of 777 (IQR: 221 to 1,906) ng/l. Patients with HFrEF were optimally treated with guideline-recommended HF therapies as illustrated by the ubiquitous use of renin-angiotensin-aldosterone blockers and beta-blockers and the high use of cardiac resynchronization therapy and implantable cardioverter-defibrillators. Online Table 1 illustrates baseline physical examination features assessing congestion states in the entire population and after subdivision according to baseline loop diuretic use, indicating that most patients had few signs of congestion at baseline.
Longitudinal spot urine concentration in the entire population
By the 30-week follow-up, a total of 1,970 individual urine samples had been collected. A total of 34 patients collected consecutive urine samples only during the first 12 weeks, as they expressed their desire to discontinue further sample collection at the first follow-up. The remaining patients collected samples for the entire 30-week follow-up. Mean Una concentration was 81.6 ± 41 mmol/l. A histogram documenting the distribution of all spot Una concentrations analyzed is shown in Figure 2. Figure 3A illustrates the repeated measurements of Una concentration with corresponding 95% confidence interval (CI) for the entire patient population. In the entire patient population, Una concentration remained stable throughout the study, as illustrated by the nonsignificant time effect (p = 0.663). However, visual inspection of all individual urinary profiles indicates interindividual differences (Online Figure 2). Post hoc t-test results of the slope and the intercept of the curve of the linear mixed model divided patients into chronically high and low sodium excretion (Figure 3B). Baseline characteristics of high and low sodium excreters are listed in Table 2, illustrating a high resemblance in baseline characteristics. Additionally, in the high sodium excreters group the slope of the individual Una concentration curve remained stable in 30 of the 39 patients (77%) and significantly dropped over time in 9 patients (23%). In the low sodium excreters group, the Una concentration remained stable in fewer patients (26 of the 41 patients or 63%), whereas 15 patients (37%) exhibited a decline in sodium output over time. Binary regression analysis, carried out to determine what clinical, biochemical, and echocardiographic factors were associated with being a high sodium excreter, are shown in Table 3, indicating that only younger age was an independent predictor of being a high sodium excreter (odds ratio [OR]: 0.91; 95% CI: 0.83 to 1.00; p = 0.045 per year).
Urinary sodium in relation to outcome
During a mean follow-up of 587 ± 54 days, a total of 21 individual patients developed the endpoint of AHF hospitalization. Follow-up was complete in all patients. Three patients who had an AHF episode subsequently died, all after the period of actively collecting urine samples. The median time to the first AHF hospitalization was 129 days (range: 71 to 248 days). Table 3 shows the baseline characteristics of the patients with or without an AHF hospitalization. Patients who had AHF hospitalization during follow-up had lower blood pressure, a lower level of hemoglobin, lower estimated glomerular filtration rate (eGFR), higher NT-proBNP, and poorer right ventricular and left ventricular systolic function and indices of higher filling pressures (E/e′ and right ventricular systolic pressure).
Figure 3C illustrates the longitudinal Una profile of patients with or without AHF hospitalization. Patients with AHF hospitalization had a lower longitudinal Una concentration compared with patients without AHF hospitalization (F[1.80] = 24.063; p < 0.001). Both the time effect and the interaction between time and development of AHF hospitalization were nonsignificant (p = 0.612 and p = 0.634, respectively), indicating that patients who develop AHF during follow-up exhibit their own statistically lower Una trajectory. Intercept analysis indicates that these curves are differentiated from one and another from the beginning (p < 0.001). In an additional sensitivity analysis including NT-proBNP and eGFR as random-effects, patients with an AHF episode still exhibited a lower Una concentration (p = 0.041). Because 34 of the patients collected samples only during the first 12 weeks, a sensitivity analysis restricted to the 46 remaining patients with a complete 30-week collection was also performed, using the same fixed and random effects in the linear mixed model. Similarly, the sensitivity analysis revealed that patients with AHF hospitalization also had a significantly lower Una concentration (F[1.46] = 5.071; p = 0.027). An additional sensitivity analysis restricted to HFrEF patients indicated similar results (F[1.218] = 20; p < 0.001), as did a sensitivity analysis restricted to patients not taking loop diuretics at baseline (F[1.203] = 11; p < 0.001). Online Table 2 illustrates baseline characteristics according to the presence of baseline loop diuretic use, whereas Online Figure 3 illustrates the longitudinal Una concentration according to baseline loop diuretic use.
A total of 16 patients developed AHF hospitalization during the period of actively collecting weekly urine samples. To further assess temporal changes in Una concentration preceding and following AHF hospitalization, the urine samples preceding and following the week of HF hospitalization are plotted in the Central Illustration. Additionally, longitudinal weight changes and changes in the VAS scale in relation to the development of AHF are also illustrated in the Central Illustration, showing that the Una concentration significantly dropped in the week preceding AHF hospitalization. This decline in Una concentration could not be identified over longer periods before decompensation (e.g., −2 weeks or −1 week). Interestingly, in comparison to Una concentration changes, early changes in weight and VAS dyspnea scoring were not apparent. Additionally, none of the patients indicated on the biometric evaluation form that their food pattern changed or became diminished (Online Table 3). After the AHF hospitalization, the Una concentration returned to a similar value as before the AHF hospitalization.
Spot samples versus continuous collection
Online Figure 4 shows an analysis of the relationship between the correlation of the averaged spot Una concentration and averaged 24-h Una concentration in patient populations with or without AHF hospitalization. Patients with AHF hospitalization had both lower spot and continuous collection Una concentrations, but the strength of the correlation did not differ from patients without an AHF episode (R-to-Z Fisher transformation p = 0.960).
The current study is the first to provide insight in the longitudinal Una profiles in stable HF patients, offering novel insights into the potential contributing mechanism of developing AHF. Main findings may include the following. 1) Stable HF patients overall exhibited a relatively stable Una concentration over time, although great interindividual differences exist. 2) Patients who developed AHF hospitalization exhibited a significantly lower Una concentration overtime. 3) Shortly before the AHF hospitalization, the Una concentration further dropped and returned to baseline after decongestive therapy. 4) A significant amount of stable HF patients excreted high concentrations of sodium without developing overt decompensation, and Una profiling may help to identify such patients.
Increasing data are pointing out that a higher Una concentration during diuretic treatment for AHF is associated with a higher odds of achieving euvolemia and a lower risk of HF readmissions and cardiovascular mortality (6,7,9). However, no data are available for chronic renal sodium excretion during a phase of clinical stability. Therefore, the present study expands the field of spot Una concentration measurements to patients with stable HF, indicating that chronic HF patients exhibit a relative stable Una profile over longer time periods, excreting relative similar amounts of Una concentrations on consecutive weeks. Interestingly, differences in LV function; use of medications, including diuretics; renal function; and residual neurohormonal activation or natriuretic peptides did not account for different patterns of Una concentration. Nevertheless, significant differences were seen between patients, with some patients excreting higher concentrations of sodium and other patients excreting lower concentrations of sodium. Intriguingly, these patients are ostensibly similar in terms of baseline characteristics, perhaps suggesting that sampling of the Una concentration can offer unique information.
Furthermore, patients who developed AHF hospital admission exhibited a significantly lower Una concentration. Additionally, the present prospective data indicate a temporal relationship between a declining Una concentration and the development of an AHF admission, which may suggest that, in the vicinity of clinical decompensation, the renal natriuretic excretion drops further. Importantly, patients did not indicate that they altered their food pattern over the weeks before decompensation. These data may be important as they provide further insight into potential mechanisms to explain why stable HF patients decompensate. Aside from pathways related to the heart (e.g., triggering factors such as ischemia, arrhythmia, or acute blood pressure changes) (14), data from patients with implantable hemodynamic monitors have illustrated that subtle rises in filling pressures precede the AHF hospitalization by numerous days (15). The present data now supplement this finding by illustrating that alterations in the renal pathway (e.g., loss of natriuretic capacity) probably contribute to and precede the development of AHF. Indeed, more sodium retention will lead to enhanced vasoreactivity and plasma volume, thereby further contributing to the occurrence of AHF (3). Interestingly, temporal changes in weight and the VAS-dyspnea score were not as apparent early on as changes in Una concentration. Indeed, it is well known that significant weight gain during the week before AHF admission only has a 6% sensitivity, underscoring the need for more sensitive tools to detect decompensation (16).
Several reasons may explain the relationship between a low Una concentration and the observed poor clinical outcome. First, it is clear that the present study cannot account for differences in the amount of ingested sodium. Therefore, it may be possible that patients who require AHF hospital admission have a lower sodium intake than patients without decompensation as a result of more adherence to instructions to sustain to a low sodium diet. More recent iterations of HF guidelines recognize that the evidence for sodium restriction is scarce, of low methodological quality, and often conflicting (12). Indeed, some observational data even suggest that low sodium intake (typically <3 g/day) may be associated with compensatory neurohormonal and sympathetic activation (17,18). Second, it can be hypothesized that patients with a chronically lower Una concentration exhibit a higher risk for AHF due to an inability to excrete sufficiently high concentrations of sodium to sustain euvolemia (4). Although some chronic sodium retention may initially be compensated for by increased nonosmotic interstitial sodium buffering, it seems that this compensatory pathway also falls short in the long run (3,19). The close temporal relationship between the further drop in Una concentration and the development of AHF may suggest an inability to buffer any additional retained sodium, leading to overt clinical decompensation. In addition, decompensation predominantly occurred in patients who already had a lower Una concentration, which may suggest the role of intrinsic renal adaptations favoring chronic tubular sodium retention (1,2). Third, it could be hypothesized that patients with an AHF episode, have mechanisms in place that result in more fluctuation in the Una concentration throughout the day, favoring a lower Una in the morning. For instance, sicker patients often manifest enhanced nocturnal diuresis, which could lower the morning Una concentration in comparison to a continuous collection. However, our data argue against this as an explanation for why HF patients with an AHF episode have a chronically lower spot Una concentration, as there was no difference in the strength between the correlation of spot versus continuous collection Una concentration in patients with or without a HF episode, perhaps arguing against more fluctuation in Una concentration due to enhanced nocturnal free water clearance in the sicker patients. However, to finitely determine whether intrinsic changes in renal sodium excretion or enhanced nocturnal free water diuresis is at the basis of the lower Una concentration we would have needed 24-h urine samples instead of Una spot samples. However, collecting longitudinally almost 2,000 urine collections is not practically feasible. Nevertheless, regardless of the mechanism that lowers Una concentration in patients who develop AHF, the Una spot samples offered novel prognostic information, even in addition to data from established biomarkers such as NT-proBNP and eGFR, potentially even detecting imminent decompensation.
Equally interesting in the present study is the finding that a large number of patients consumed and excreted high concentrations of sodium. This may indicate that a subset of stable HF patients could safely ingest large quantities of sodium because they do not develop a subsequent AHF admission. Although 1 observational study in healthy individuals linked a higher Una output to a higher incidence of new onset HF, this association was not statistically significant after adjusting for blood pressure (20). Our HF patients, under optimal neurohormonal blockade, did not exhibit hypertension. It is therefore questionable whether sodium restriction is truly necessary in all HF patients.
Although the present study is hypothesis generating, the data indicate that longitudinal assessment of Una is feasible and may identify different subsets of stable patients. Although Una concentrations were analyzed in a laboratory setting, self-measurement of Una or chloride concentration is possible by using a urinary dipstick method (21). If additional studies confirm our findings, Una profiling may become a tool to empower HF patients to predict imminent decompensation or to individualize dietary patterns or the diuretic regimen.
First, this is a small study, and therefore, all results should be interpreted as exploratory and hypothesis generating. Nevertheless, the sample size of the study should be interpreted in the face of the intensive study protocol. Second, spot sodium values were used, as longitudinally collecting a large number of 24-h urine samples is impractical. However, there is increasing interest in spot Una concentration, and this method can easily be integrated into clinical practice. Third, there was some subject dropout in our study (20%), which relates to the intensive study protocol. Fourth, differences in sodium intake cannot be accounted for. Fifth, dichotomizing the individual Una concentration curve into high and low sodium excreters is a somewhat artificial statistical method. However, the main goal of this analysis was to determine whether different urinary phenotypes exist that may not be readily apparent when looking at baseline characteristics. Due to the small sample size, division into more groups (e.g., an intermediate group) was not possible. Finally, only a sample number of HFpEF patients were enrolled, and it is clearly not powered to assess this cohort. However, a sensitivity analysis restricted to the HFrEF cohort did demonstrate the incremental information of Una concentration.
Una concentration remains relatively consistent over time in stable, mildly symptomatic HF patients. Patients who require an HF admission exhibit a chronically lower Una concentration, which further drops just before decompensation.
COMPETENCY IN MEDICAL KNOWLEDGE: The burden of HF readmission remains high in clinical practice. Strategies that help to understand the pathophysiology of the development of acute decompensation are necessary. The present analysis demonstrates that patients who are at risk for development of AHF chronically excrete a lower amount of sodium. Additionally, a further drop in sodium excretion precedes the development of AHF. Urinary sodium profiling may help to detect patients at risk for AHF and possibly imminent decompensation.
TRANSLATIONAL OUTLOOK: Measurement of urinary sodium in a laboratory setting is inexpensive, readily available, and may offer novel information on HF patients. However, self-measurement of urinary sodium or chloride is possible by using a dipstick method. Research is necessary to determine if this may be a tool to empower HF patients. Additionally, research is necessary to determine if tailored therapy to enhance sodium excretion is associated with a lower risk for AHF.
Dr. Martens is supported by a doctoral fellowship from Research Foundation-Flanders grant 1127917N. Drs. Martens and Mullens perform research for the Limburg Clinical Research Program UHasselt-ZOL-Jessa, supported by the Limburg Sterk Merk Foundation, Hasselt University, Ziekenhuis Oost-Limburg, and Jessa Hospital. Dr. Tang has consulted for Sequana Medical Inc. and MyoKardia Inc. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- acute heart failure
- confidence interval
- estimated glomerular filtration rate
- heart failure
- heart failure with reduced ejection fraction
- heart failure with preserved ejection fraction
- N-terminal pro–B-type natriuretic peptide
- odds ratio
- urinary sodium
- visual analog scale
- Received January 1, 2019.
- Revision received February 14, 2019.
- Accepted February 20, 2019.
- 2019 American College of Cardiology Foundation
- Mullens W.,
- Verbrugge F.H.,
- Nijst P.,
- Tang W.H.W.
- Nijst P.,
- Verbrugge F.H.,
- Grieten L.,
- et al.
- McKie P.M.,
- Schirger J.A.,
- Costello-Boerrigter L.C.,
- et al.
- Nijst P.,
- Martens P.,
- Dupont M.,
- Tang W.H.W.,
- Mullens W.
- Brinkley D.M. Jr..,
- Burpee L.J.,
- Chaudhry S.P.,
- et al.
- Ferreira J.P.,
- Girerd N.,
- Medeiros P.B.,
- et al.
- Martens P.,
- Mullens W.
- Testani J.M.,
- Hanberg J.S.,
- Cheng S.,
- et al.
- Verbrugge F.H.,
- Nijst P.,
- Dupont M.,
- Penders J.,
- Tang W.H.,
- Mullens W.
- Ponikowski P.,
- Voors A.A.,
- Anker S.D.,
- et al.
- Van Aelst L.N.L.,
- Arrigo M.,
- Placido R.,
- et al.
- Adamson P.B.,
- Magalski A.,
- Braunschweig F.,
- et al.
- Doukky R.,
- Avery E.,
- Mangla A.,
- et al.
- Nijst P.,
- Olinevich M.,
- Hilkens P.,
- et al.
- Verbrugge F.H.,
- Martens P.,
- Boonen L.,
- et al.