Int Urol Nephrol DOI 10.1007/s11255-014-0721-2
NEPHROLOGY - ORIGINAL PAPER
Characteristics, therapies, and factors influencing outcomes of hospitalized hypernatremic geriatric patients Muhammad R. Toor • Anjali Singla • Maria V. DeVita • Jordan L. Rosenstock Michael F. Michelis
•
Received: 6 November 2013 / Accepted: 21 April 2014 Ó Springer Science+Business Media Dordrecht 2014
Abstract Purpose Hypernatremia is a common electrolyte disorder associated with adverse outcomes such as increased length of stay and mortality due to a variety of factors. Our aim was to investigate known factors as well as other variables which we had identified in hospitalized hypernatremic geriatric patients and their relationship to patient outcomes. Methods A retrospective chart review of all adult hospitalized patients in a 4-month period with a serum sodium level [150 mmol/L was performed. Factors evaluated included use of a nephrology consultation, certain urine laboratory measures, fluids employed, rate of correction, and patient’s level of care setting. Outcome measures included length of stay and mortality. Results The patient mortality rate was 52 %. Mean age was 79.6 years (n = 33), and mean initial sodium level was 152.6 mmol/L. Plasma and urine osmolality, and urine sodium concentration were checked in less than 25 % of patients. Fifteen of 18 patients in the ICU expired, whereas only 2 of 15 patients not in the ICU expired (p \ 0.0004, OR 32.50, CI 95 % (4.68–225.54)). Of the 23 patients
(70 %) who had their serum sodium level corrected, 11 were corrected in B3 days and 12 in [3 days, but this difference did not affect mortality rate (45 vs. 50 %, p = 0.99). The mortality rate was similar (60 %, p = 0.52) for those whose serum sodium level never corrected suggesting that correction did not influence outcomes. The fluids chosen for therapy of the hypernatremia were appropriate to the patients volume status. Five of 15 patients who received a nephrology consultation survived, while 11 of 18 patients without a nephrology consultation survived (p = 0.12). The mean length of stay was 25.0 ± 23.9 days and no different for those who expired versus those who survived (25.2 ± 21.2 vs. 24.8 ± 25.9 days, p = 0.96). Conclusions Hypernatremia is associated with a poor prognosis, and outcomes are still disappointing despite appropriate rates of correction, intensive monitoring, and the involvement of a nephrologist. Strategies directed at avoidance of the development of hypernatremia and attention to concomitant disease may provide significant patient benefit. Keywords Hypernatremia Fluid therapies Urine electrolytes Geriatric patients
M. R. Toor (&) A. Singla M. V. DeVita J. L. Rosenstock M. F. Michelis Nephrology, Lenox Hill Hospital, 100 East 77th Street, New York, NY 10075, USA e-mail: [email protected] A. Singla e-mail: [email protected] M. V. DeVita e-mail: [email protected] J. L. Rosenstock e-mail: [email protected] M. F. Michelis e-mail: [email protected]
Introduction Hypernatremia is a common electrolyte disorder and is associated with adverse outcomes such as increased hospital length of stay and increased mortality [1–3]. The incidence of hypernatremia has been reported to be 5 % of all hospitalized patients with mortality rates ranging from 37 to 66 % [4, 5]. Prior studies have suggested a variety of factors that may result in undesirable outcomes. In a prospective cohort
123
Int Urol Nephrol
study analyzing 103 patients with a serum sodium concentration of 150 mmol/L or greater present either on hospital admission or during hospitalization, Palevsky et al. concluded that hospital acquired hypernatremia was primarily iatrogenic. Their results suggested that patients with predictable increased water losses and impaired thirst and/ or restricted free water intake were given either an inadequate or inappropriate prescription of fluids. They also concluded that treatment for hypernatremia is often delayed [3]. In a more recent study evaluating whether physicians follow the recommended guideline for the rate of correction of hypernatremia, the data suggested that correction of hypernatremia was often slow in patients with severe hypernatremia (serum sodium C155 mEq/L) [4]. Further, the slow rate of correction was significantly associated with 30-day patient mortality [4]. Considering the severity of outcomes related to hypernatremia as well as the factors previously implicated to be associated with such outcomes, our aim was to investigate whether other known variables which we had identified in hospitalized hypernatremic geriatric patients also influenced these outcomes. We evaluated the location of therapy in the hospital, the use of nephrology consultation, whether plasma and urine laboratory studies were obtained, and types of fluid employed for treatment, in addition to more traditional factors such as the rate of correction of serum sodium and the mobility and cognitive status of our patients.
Methods We performed a retrospective cohort study of patients with hypernatremia from May 31, 2012 to September 30, 2012, at Lenox Hill Hospital (an acute care hospital located in New York, New York). We identified hospitalized patients over the age of 18 with a serum sodium level [150 mmol/ L using the hospital’s clinical database system. All data were obtained from this database and patient medical records. The following characteristics for each patient were recorded: age, gender, symptoms related to hypernatremia (yes vs. no), receiving tube feeding (yes vs. no), principal medical diagnosis, origin (e.g., home, nursing home, etc.), bed confinement (yes vs. no), history of diabetes mellitus (yes vs. no), history of lithium therapy (yes vs. no), nephrology consultation obtained (yes vs. no), plasma osmolality level measured (yes vs. no), urine osmolality level measured (yes vs. no), urine sodium level measured (yes vs. no), initial ward setting, transfer to another ward setting (yes vs. no), whether the elevated sodium level was corrected to B145 mmol/L (yes vs. no), number of days to correction, decision to use fluid therapy (yes vs. no), and
123
type of initial fluid therapy employed. The primary outcomes included length of stay, inpatient mortality, and discharge setting. Exclusion criteria consisted of confirmed or suspected laboratory contamination/error and if hypertonic saline had been used as part of fluid therapy. Statistical data were evaluated using independent t test, Chi-square analysis, and Fisher’s exact test. A p value less than 0.05 was considered significant. All data are expressed as mean ± SD. The study was approved by the Institutional Review Board of North Shore-Long Island Jewish Health System.
Results Of 38 patients with hypernatremia, 5 were excluded because of data deficiencies. Of the remaining 33 patients studied, the mean age was 79.6 ± 11.03 years. Twelve patients were male, and 21 were female with mean ages of 78.9 and 80.0 years, respectively. The mean initial serum sodium level was 152.6 ± 2.16 mmol/L (range 151–160 mmol/L). Twenty-six patients demonstrated CNS symptomatology, which included altered mental status, coma, and/or seizures, while 7 did not. Twenty patients presented from home, 12 from an advanced care facility, and 1 patient was transferred from another hospital. Patient characteristics are summarized on Table 1. The principle medical diagnoses were categorized as follows: heart-related disease (n = 9), pulmonary-related disease (n = 8), sepsis (n = 7), malignancy (n = 3), urinary tract infection (n = 2), peripheral vascular disease (n = 1), cerebrovascular accident (n = 1), and other (n = 2) (Fig. 1). Twenty-six out of 33 patients were confined to bed indicating their impaired status. Patients were evenly divided as to whether or not they were receiving tube feedings (16 vs. 17 patients). Eight patients had a
Table 1 Patient characteristics Age (%) \75
10 (30.3 %)
75–85 C86
9 (27.3 %) 14 (42.4 %)
Gender (%) Male Female CNS symptomatology (%)
12 (36.4 %) 21 (63.6 %) 26 (78.8 %)
Intake setting (%) Home
20 (60.6 %)
Advanced care facility
12 (36.4 %)
Transfer from outside hospital
1 (3.0 %)
Int Urol Nephrol Table 3 Multivariable logistic regression analysis: variables versus mortality
Fig. 1 Principal medical diagnosis for study patients
Table 2 Baseline clinical and laboratory variables Variable
n
Minimum
Maximum
Age (years)
33
79.6 ± 11.03
56
99
Mean arterial pressure (mmHg)
33
85.7 ± 16.39
51
131
Initial serum sodium (mmol/L)
33
152.6 ± 2.16
151
160
Initial serum glucose (mg/dL)
33
135.4 ± 49.88
65
277
Initial BUN (mg/dL)
33
45.5 ± 26.88
Initial serum creatinine (mg/dL)
33
1.4 ± 1.06
5
52.2 ± 39.28
Urine sodium concentration (mmol/L)
Mean ± SD
6
117
0.25
4.7
21
119
history of diabetes mellitus, 6 of whom were insulin dependent. No patients had a history of lithium therapy. The overall mortality rate was 52 % (17 died and 16 survived). Plasma osmolality was checked in only 1 patient (3 %), urine osmolality in 7 patients (21 %), and urine sodium in 5 patients (15 %). Baseline clinical and laboratory variables are shown in Table 2. Fifteen of 18 patients admitted to or transferred to the ICU expired, whereas only 2 of 15 patients not in the ICU expired (p \ 0.0004, OR 32.50, CI 95 % (4.68–225.54)). Mean serum sodium of patients in an ICU setting was 152.5 ± 1.3 mmol/L (range 151–154 mmol/L), compared to 152.7 ± 2.9 mmol/L for those in a non-ICU setting (range 151–160 mmol/L), (p = 0.80). Twenty-three (70 %) patients had their serum sodium level corrected. Eleven patients were promptly corrected in B3 days and 12 in [3 days, but this difference did not affect survival (5 died, 6 survived; mortality rate = 45 % vs. 6 died, 6 survived; mortality rate = 50 %, p = 0.99). Of the 10 patients who did not have their serum sodium level corrected, 6 died and 4 remained alive and the
Variable
p value
Odds ratio
Age ([75 years)
0.90
1.10
Gender (Female)
0.90
1.10
Ward setting (ICU) Sodium not corrected
0.0004 0.52
Nephrology consultation not obtained
95 % CI 0.27–4.55 0.27–4.55
32.50 1.64
4.68–225.54 0.36–7.38
0.12
0.32
0.08—1.33
Diabetes mellitus
0.71
1.81
0.35–9.24
Initial serum sodium (C155 mmol/L)
0.28
0.27
0.03–2.92
CNS symptomatology
0.61
1.56
0.29–8.38
Initial serum creatinine (C1.5 mg/dL)
0.56
1.54
0.37–6.45
mortality rate of 60 % was not different from that of patients whose serum sodium was corrected (p = 0.52). Of the 33 patients studied, 15 received a consultation by a nephrologist and of these, 5 survived and 10 died. Of the 18 patients without nephrology consultation, outcomes were not different, 11 lived and 7 died (p = 0.12). Thirteen of the 15 nephrology consultations obtained were within 1 day in regard to the recognition of the patients’ hypernatremic status. Eight of these patients died (62 %). In the 2 patients in which a nephrology consult was obtained[24 h after the recognition of hypernatremia, both died. Analysis of patient data also revealed that presence of diabetes mellitus, initial serum sodium concentration C155 mmol/ L, and CNS symptomatology did not influence mortality. Multivariable logistic regression analysis of variables versus mortality is depicted in Table 3. Thirty-one (94 %) patients received various IVF and/or oral fluids, including 0.9 % normal saline (NS), 0.45 % NS, D5-0.45 % NS, 0.3 % NS, D5, lactated ringers (LR), and oral water per os or via tube. The choice of fluid appeared appropriate to the clinical status with hypovolemic patients receiving normal saline, 0.45 % NS for uncertain volume status and water for those patients clearly only water deficient (Fig. 2). The mean length of stay for all patients was 25.0 ± 23.9 days and no different for those who expired versus those who remained alive (25.2 ± 21.2 vs. 24.8 ± 25.9 days, p = 0.96). Moreover, the mean length of stay was no different if a nephrology consultation was obtained (21.6 ± 20.8 vs. 20.1 ± 21.8 days, p = 0.84). Of the patients who presented from home (n = 20), 12 died, 5 returned to home, and 3 were sent to an advanced care facility. Of the patients who presented from an advanced care facility (n = 12), 4 died, zero went home, 5 returned to an advanced care facility, and 3 were sent to
123
Int Urol Nephrol
Fig. 2 Type of fluid selected to correct hypernatremia
Fig. 3 Discharge setting. Fifty-two percent of patients expired
hospice. The only patient who was transferred from an outside hospital died. Figure 3 summarizes the discharge settings of our patients.
Discussion In this retrospective study analyzing outcomes of hospitalized patients with hypernatremia, the mortality rate was 52 %, which is consistent with prior studies, with approximately half of the survivors being discharged to a facility with specialized services indicating the fragile nature of these patients. Hypernatremia was associated with advanced age, and the mean length of stay did not appear to be different for survivors compared to nonsurvivors. Cardiac-related disease, pulmonary-related disease, and sepsis comprised the majority of principle medical diagnoses related to hypernatremia. The mortality rate was found to be high despite prompt rate of correction, involvement of a medical specialist, and intensive care monitoring. Patients were more likely to expire whether they were admitted to or transferred to the ICU compared to receiving all of their care in a telemetry unit or in a general medical ward. The serum sodium concentration is usually maintained within a narrow range of 138–142 mmol/L despite great
123
variation in water intake among individuals. This is accomplished by regulating the balance of water through responses to osmoreceptors and baroreceptors that regulate the secretion of antidiuretic hormone (ADH) from the hypothalamus [6]. The fact that in our study, hypernatremia was associated with an elderly population is not surprising considering the physiologic changes associated with aging which predispose to hypernatremia as well as certain risk factors that increase their susceptibility to the condition. With advancing age, the ability to concentrate the urine diminishes and osmotic stimulation of thirst also diminishes, even though the mechanisms associated with the release of ADH are maintained [7–9]. It is speculated that the reduced ability to maximally concentrate the urine with aging may be related to decreased activity of many of the key transport proteins (aquaporin (AQP) 2, serine-256— phosphorylated AQP2, AQP3, NKCC2/BSC1, urea transporter genes UT-A1 and UT-B, and the V2-receptor) that contribute to medullary hypertonicity and water transport [10, 11]. Our study suggested that those patients managed in the ICU had a higher mortality rate (83 %) than those in nonintensive care areas, despite similar mean serum sodium levels. An explanation for this finding may be that such patients exhibited medical conditions which were more severe compared to those who were felt to be manageable in a non-ICU setting. In a study performed by Hoorn et al. [12] investigating the risk factors and mechanisms of hypernatremia in critically ill patients, they noted that the patients with ICU acquired hypernatremia were generally sicker on admission to the ICU, which was reflected by a higher Acute Physiology And Chronic Health Evaluation II (APACHE II) score and by a lower Glasgow Coma Score (GCS). The forty-five percent mortality observed in our study in patients who had hypernatremia corrected in 72 h or less is in contrast to earlier data. In a study by Alshayeb et al. evaluating the effect of the rate of correction of severe hypernatremia (defined as serum sodium C155 mEq/L) on the mortality of hospitalized patients, they found that patients whose hypernatremia was corrected within 72 h had a lower mortality (10 vs. 39 %, p = 0.0048) and were more likely to be managed in the ICU (89 vs. 64 %, p = 0.0096). This was observed in 107 patients who survived beyond 72 h [4]. Our study indicated that there was no benefit in regard to mortality for patients with hypernatremia corrected within 72 h. Factors that may influence the effect of prompt rate of correction may relate to a beneficial effect on more acutely hypernatremic patients as well as the severity of the patient’s condition. Indeed, our data revealed that survival was not improved whether or not the serum sodium was corrected. The diseases associated with hypernatremia in our patients included conditions
Int Urol Nephrol
with multi-organ failure in the setting of cardiac disease, advanced pneumonia, and sepsis indicating that disease severity was the major influence regarding outcome. Since hypernatremia is not a uniform disorder, treatment is often a challenging task. Current recommendations for the treatment for hypernatremia in elderly patients include calculation of the free water deficit [0.5 9 weight in kg] 9 [1 - (140/serum sodium)] and the correction of half the deficit in the first 24 h [13, 14]. Since ongoing losses of hypotonic fluids (i.e., gastrointestinal loss, pyrexia) may contribute, these losses should also be added to the total water deficit calculation. If there is concern that an osmotic diuresis is occurring, then measurements of plasma sodium, urine volume, and urine electrolyte concentrations (sodium and potassium) are used for the determination of electrolytefree water clearance (CeH2O = VU 9 (1 - ([Na]U ? [K]U)/[Na]S)) so that ongoing free water loss can also be added [15]. If the dysnatremia is acute over a few hours (e.g., iatrogenic sodium load), then it is appropriate to correct at a rate of 1–2 mmol/L/h until symptoms subside. If the change in sodium concentration occurred over a longer or unknown duration, the correction rate can be 0.5 mmol/L/h with a target of less than 12 mmol/L in a 24-h period to avoid cerebral edema and neurotoxicity [1]. Cerebral edema may actually be more of a concern in young children because of small skull volume [16, 17]. A hypotonic solution, such as pure water, D5, 0.3 % NS, or 0.45 % NS, should be used to decrease serum sodium. The preferred route of water administration is oral, per mouth, or through a feeding tube if necessary. If parenteral D5 is used, attention to correction of hyperglycemia is important. Normal saline is an appropriate initial therapy option if the patient is hypovolemic and must be fluid resuscitated [13]. Also, diuretics may be needed if intravascular volume excess develops or to assist with the excretion of hypertonic fluid. In summary, it is essential to diagnose the condition causing the hypernatremia along with assessing the patient’s volume status, estimating the water requirements, selecting a proper fluid type with an appropriate rate of correction, and monitoring the patient and laboratory values closely during therapy [18]. Our study differed from prior studies analyzing outcomes in patients with hypernatremia in several ways. First, we investigated certain variables in hospitalized hypernatremic geriatric patients which have not been previously evaluated. These included whether or not a nephrologist was involved in the case and the origin of the patient prior to admission. An interesting finding of our study was that the mortality rate observed in patients receiving a nephrology consultation was substantial (67 %; 5 alive, 10 died) compared to those who did not receive one
(39 %; 11 alive, 7 died) (p = 0.12). It seemed, therefore, that nephrology consultation did not appreciably influence patient outcomes. Review of patient data revealed that the timing of nephrology consultation often occurred early in the patient’s hospital course, but this also did not appear to affect the results noted. Secondly, our study quantitated the use of diagnostic urine laboratory measurements in our patients. Less than 25 % of patients had any urine studies performed which we feel would have helped in the correct evaluation of urinary concentrating ability and volume status [13]. Patients with low urine sodium levels and a concentrated urine could be classified to be volume depleted requiring volume resuscitation, while those with urine osmolalities suggesting a failure of urinary concentration would be those who required primarily water replacement. The observations regarding the limited use of plasma and urine laboratory studies are similar to those reported by Huda et al. [19] in a report describing hyponatremic patients and may represent a general lack of appreciation of the utility of these studies by clinicians. Our results indicate that hypernatremia continues to be associated with a high mortality rate. Prevention of hypernatremia may be more important than some of the therapeutic variables cited above. Our study revealed that approximately 40 % of patients presented from a facility other than home, most likely because the patient was unable to perform self-care. Since the geriatric population is already predisposed to the development of hypernatremia due to physiologic changes related to aging, certain measures could and should be taken for preventing hypernatremia, especially in those individuals that require supervision and an extra level of care (i.e., individuals in nursing homes). One such measure could be to obtain blood chemistries at more frequent intervals, particularly in patients receiving tube feeding or with limited mobility. This would be performed to not only assess the serum sodium level but also assess the serum urea and serum glucose levels since an osmotic diuresis may ensue from high serum concentration of these substances, thereby contributing to hypernatremia [20]. Another measure that could be employed would be the recording of input/output data more frequently with special attention to evaluation of fluid, especially water, intake. In conclusion, hypernatremia is associated with a poor prognosis, and outcomes are still disappointing despite appropriate rates of correction, intensive monitoring, and the involvement of a medical specialist. Implementation of strategies directed at the avoidance of hypernatremia, as well as vigilance regarding the development of new or the worsening of concomitant disorders, may provide patient benefit.
123
Int Urol Nephrol Conflict of interest The authors of this manuscript declare that they have no conflict of interest.
References 1. Adrogue´ HJ, Madias NE (2000) Hypernatremia. N Engl J Med 342(20):1493–1499 2. Lindner G, Funk GC, Schwarz C, Kneidinger N, Kaider A, Schneeweiss B, Druml W (2007) Hypernatremia in the critically ill is an independent risk factor for mortality. Am J Kidney Dis 50:952–957. doi:10.1053/j.ajkd.2007.08.016 3. Palevsky PM, Bhagrath R, Greenberg A (1996) Hypernatremia in hospitalized patients. Ann Intern Med 124:197–203. doi:10.7326/ 0003-4819-124-2-199601150-00002 4. Alshayeb HM, Showkat A, Babar F, Mangold T, Wall BM (2011) Severe hypernatremia correction rate and mortality in hospitalized patients. Am J Med Sci 341:356–360. doi:10.1097/MAJ. 0b013e31820a3a90 5. Mandal AK, Saklayen MG, Hillman NM, Markert RJ (1997) Predictive factors for high mortality in hypernatremia patients. Am J Emerg Med 15:130–132. doi:10.1016/S07356757(97)90082-6 6. Kumar S, Berl T (1998) Sodium. Lancet 352:220–228. doi:10. 1016/S0140673697121699 7. Rowe JW, Shock NW, DeFronzo R (1976) The influence of age on the renal response to water deprivation in man. Nephron 17:270–278. doi:10.1159/000180731 8. Kenney WL, Chiu P (2001) Influence of age on thirst and fluid intake. Med Sci Sports Exerc 33:1524–1532. doi:10.1097/ 00005768-200109000-00016 9. Rose BD, Post TW (2001) Hyperosmolal states–hypernatremia. In: Clinical physiology of acid-base and electrolyte disorders, 5th edn. McGraw-Hill, New York, pp 746–793
123
10. Sands JM (2009) Urinary concentration and dilution in the aging kidney. Semin Nephrol 29:579–586. doi:10.1016/j.semnephrol. 2009.03.008 11. Zhou XJ, Saxena R, Liu Z, Vaziri ND, Silva FG (2008) Renal senescence in 2008: progress and challenges. Int Urol Nephrol 40:823–839. doi:10.1007/s11255-008-9405-0 12. Hoorn EJ, Betjes MG, Weigel J, Zietse R (2008) Hypernatraemia in critically ill patients: too little water and too much salt. Nephrol Dial Transplant 23:1562–1568. doi:10.1093/ndt/gfm831 13. Michelis MF (2009) Chapter 16: disorders of serum sodium concentration in the elderly patient. In: Geriatric nephrology curriculum, American Society of Nephrology pp 1–6 14. Adrogue HJ, Madias NE (1997) Aiding fluid prescription for the dysnatremias. Intensive Care Med 23:309–316. doi:10.1007/ s001340050333 15. Popli S, Tzamaloukas AH, Ing TS (2013) Osmotic diuresisinduced hypernatremia: better explained by solute-free water clearance or electrolyte-free water clearance? Int Urol Nephrol (Epub January 2013). doi 10.1007/s11255-012-0353-3 16. Ayus JC, Brennan S (1995) Hypernatremia. In: Arieff AI, DeFronzo RA (eds) Fluid, electrolyte, and acid-base disorders, 2nd edn. Churchill Livingstone, New York, pp 304–317 17. Xenos C, Sgouros S, Natarajan K (2002) Ventricular volume change in childhood. J Neurosurg 97:584–590 18. Al-Absi A, Gosmanova EO, Wall BM (2012) A clinical approach to the treatment of chronic hypernatremia. Am J Kidney Dis 60:1032–1038. doi:10.1053/j.ajkd.2012.06.025 19. Huda MB, Boyd A, Skagen K, Wile D, Van Heyningen C, Watson I, Wong S, Gill G (2006) Investigation and management of severe hyponatraemia in a hospital setting. Postgrad Med J 82:216–219. doi:10.1136/pmj.2005.036947 20. Lindner G, Schwarz C, Funk GC (2012) Osmotic diuresis due to urea as the cause of hypernatraemia in critically ill patients. Nephrol Dial Transplant 27:962–967. doi:10.1093/ndt/gfr428
NEPHROLOGY - ORIGINAL PAPER
Characteristics, therapies, and factors influencing outcomes of hospitalized hypernatremic geriatric patients Muhammad R. Toor • Anjali Singla • Maria V. DeVita • Jordan L. Rosenstock Michael F. Michelis
•
Received: 6 November 2013 / Accepted: 21 April 2014 Ó Springer Science+Business Media Dordrecht 2014
Abstract Purpose Hypernatremia is a common electrolyte disorder associated with adverse outcomes such as increased length of stay and mortality due to a variety of factors. Our aim was to investigate known factors as well as other variables which we had identified in hospitalized hypernatremic geriatric patients and their relationship to patient outcomes. Methods A retrospective chart review of all adult hospitalized patients in a 4-month period with a serum sodium level [150 mmol/L was performed. Factors evaluated included use of a nephrology consultation, certain urine laboratory measures, fluids employed, rate of correction, and patient’s level of care setting. Outcome measures included length of stay and mortality. Results The patient mortality rate was 52 %. Mean age was 79.6 years (n = 33), and mean initial sodium level was 152.6 mmol/L. Plasma and urine osmolality, and urine sodium concentration were checked in less than 25 % of patients. Fifteen of 18 patients in the ICU expired, whereas only 2 of 15 patients not in the ICU expired (p \ 0.0004, OR 32.50, CI 95 % (4.68–225.54)). Of the 23 patients
(70 %) who had their serum sodium level corrected, 11 were corrected in B3 days and 12 in [3 days, but this difference did not affect mortality rate (45 vs. 50 %, p = 0.99). The mortality rate was similar (60 %, p = 0.52) for those whose serum sodium level never corrected suggesting that correction did not influence outcomes. The fluids chosen for therapy of the hypernatremia were appropriate to the patients volume status. Five of 15 patients who received a nephrology consultation survived, while 11 of 18 patients without a nephrology consultation survived (p = 0.12). The mean length of stay was 25.0 ± 23.9 days and no different for those who expired versus those who survived (25.2 ± 21.2 vs. 24.8 ± 25.9 days, p = 0.96). Conclusions Hypernatremia is associated with a poor prognosis, and outcomes are still disappointing despite appropriate rates of correction, intensive monitoring, and the involvement of a nephrologist. Strategies directed at avoidance of the development of hypernatremia and attention to concomitant disease may provide significant patient benefit. Keywords Hypernatremia Fluid therapies Urine electrolytes Geriatric patients
M. R. Toor (&) A. Singla M. V. DeVita J. L. Rosenstock M. F. Michelis Nephrology, Lenox Hill Hospital, 100 East 77th Street, New York, NY 10075, USA e-mail: [email protected] A. Singla e-mail: [email protected] M. V. DeVita e-mail: [email protected] J. L. Rosenstock e-mail: [email protected] M. F. Michelis e-mail: [email protected]
Introduction Hypernatremia is a common electrolyte disorder and is associated with adverse outcomes such as increased hospital length of stay and increased mortality [1–3]. The incidence of hypernatremia has been reported to be 5 % of all hospitalized patients with mortality rates ranging from 37 to 66 % [4, 5]. Prior studies have suggested a variety of factors that may result in undesirable outcomes. In a prospective cohort
123
Int Urol Nephrol
study analyzing 103 patients with a serum sodium concentration of 150 mmol/L or greater present either on hospital admission or during hospitalization, Palevsky et al. concluded that hospital acquired hypernatremia was primarily iatrogenic. Their results suggested that patients with predictable increased water losses and impaired thirst and/ or restricted free water intake were given either an inadequate or inappropriate prescription of fluids. They also concluded that treatment for hypernatremia is often delayed [3]. In a more recent study evaluating whether physicians follow the recommended guideline for the rate of correction of hypernatremia, the data suggested that correction of hypernatremia was often slow in patients with severe hypernatremia (serum sodium C155 mEq/L) [4]. Further, the slow rate of correction was significantly associated with 30-day patient mortality [4]. Considering the severity of outcomes related to hypernatremia as well as the factors previously implicated to be associated with such outcomes, our aim was to investigate whether other known variables which we had identified in hospitalized hypernatremic geriatric patients also influenced these outcomes. We evaluated the location of therapy in the hospital, the use of nephrology consultation, whether plasma and urine laboratory studies were obtained, and types of fluid employed for treatment, in addition to more traditional factors such as the rate of correction of serum sodium and the mobility and cognitive status of our patients.
Methods We performed a retrospective cohort study of patients with hypernatremia from May 31, 2012 to September 30, 2012, at Lenox Hill Hospital (an acute care hospital located in New York, New York). We identified hospitalized patients over the age of 18 with a serum sodium level [150 mmol/ L using the hospital’s clinical database system. All data were obtained from this database and patient medical records. The following characteristics for each patient were recorded: age, gender, symptoms related to hypernatremia (yes vs. no), receiving tube feeding (yes vs. no), principal medical diagnosis, origin (e.g., home, nursing home, etc.), bed confinement (yes vs. no), history of diabetes mellitus (yes vs. no), history of lithium therapy (yes vs. no), nephrology consultation obtained (yes vs. no), plasma osmolality level measured (yes vs. no), urine osmolality level measured (yes vs. no), urine sodium level measured (yes vs. no), initial ward setting, transfer to another ward setting (yes vs. no), whether the elevated sodium level was corrected to B145 mmol/L (yes vs. no), number of days to correction, decision to use fluid therapy (yes vs. no), and
123
type of initial fluid therapy employed. The primary outcomes included length of stay, inpatient mortality, and discharge setting. Exclusion criteria consisted of confirmed or suspected laboratory contamination/error and if hypertonic saline had been used as part of fluid therapy. Statistical data were evaluated using independent t test, Chi-square analysis, and Fisher’s exact test. A p value less than 0.05 was considered significant. All data are expressed as mean ± SD. The study was approved by the Institutional Review Board of North Shore-Long Island Jewish Health System.
Results Of 38 patients with hypernatremia, 5 were excluded because of data deficiencies. Of the remaining 33 patients studied, the mean age was 79.6 ± 11.03 years. Twelve patients were male, and 21 were female with mean ages of 78.9 and 80.0 years, respectively. The mean initial serum sodium level was 152.6 ± 2.16 mmol/L (range 151–160 mmol/L). Twenty-six patients demonstrated CNS symptomatology, which included altered mental status, coma, and/or seizures, while 7 did not. Twenty patients presented from home, 12 from an advanced care facility, and 1 patient was transferred from another hospital. Patient characteristics are summarized on Table 1. The principle medical diagnoses were categorized as follows: heart-related disease (n = 9), pulmonary-related disease (n = 8), sepsis (n = 7), malignancy (n = 3), urinary tract infection (n = 2), peripheral vascular disease (n = 1), cerebrovascular accident (n = 1), and other (n = 2) (Fig. 1). Twenty-six out of 33 patients were confined to bed indicating their impaired status. Patients were evenly divided as to whether or not they were receiving tube feedings (16 vs. 17 patients). Eight patients had a
Table 1 Patient characteristics Age (%) \75
10 (30.3 %)
75–85 C86
9 (27.3 %) 14 (42.4 %)
Gender (%) Male Female CNS symptomatology (%)
12 (36.4 %) 21 (63.6 %) 26 (78.8 %)
Intake setting (%) Home
20 (60.6 %)
Advanced care facility
12 (36.4 %)
Transfer from outside hospital
1 (3.0 %)
Int Urol Nephrol Table 3 Multivariable logistic regression analysis: variables versus mortality
Fig. 1 Principal medical diagnosis for study patients
Table 2 Baseline clinical and laboratory variables Variable
n
Minimum
Maximum
Age (years)
33
79.6 ± 11.03
56
99
Mean arterial pressure (mmHg)
33
85.7 ± 16.39
51
131
Initial serum sodium (mmol/L)
33
152.6 ± 2.16
151
160
Initial serum glucose (mg/dL)
33
135.4 ± 49.88
65
277
Initial BUN (mg/dL)
33
45.5 ± 26.88
Initial serum creatinine (mg/dL)
33
1.4 ± 1.06
5
52.2 ± 39.28
Urine sodium concentration (mmol/L)
Mean ± SD
6
117
0.25
4.7
21
119
history of diabetes mellitus, 6 of whom were insulin dependent. No patients had a history of lithium therapy. The overall mortality rate was 52 % (17 died and 16 survived). Plasma osmolality was checked in only 1 patient (3 %), urine osmolality in 7 patients (21 %), and urine sodium in 5 patients (15 %). Baseline clinical and laboratory variables are shown in Table 2. Fifteen of 18 patients admitted to or transferred to the ICU expired, whereas only 2 of 15 patients not in the ICU expired (p \ 0.0004, OR 32.50, CI 95 % (4.68–225.54)). Mean serum sodium of patients in an ICU setting was 152.5 ± 1.3 mmol/L (range 151–154 mmol/L), compared to 152.7 ± 2.9 mmol/L for those in a non-ICU setting (range 151–160 mmol/L), (p = 0.80). Twenty-three (70 %) patients had their serum sodium level corrected. Eleven patients were promptly corrected in B3 days and 12 in [3 days, but this difference did not affect survival (5 died, 6 survived; mortality rate = 45 % vs. 6 died, 6 survived; mortality rate = 50 %, p = 0.99). Of the 10 patients who did not have their serum sodium level corrected, 6 died and 4 remained alive and the
Variable
p value
Odds ratio
Age ([75 years)
0.90
1.10
Gender (Female)
0.90
1.10
Ward setting (ICU) Sodium not corrected
0.0004 0.52
Nephrology consultation not obtained
95 % CI 0.27–4.55 0.27–4.55
32.50 1.64
4.68–225.54 0.36–7.38
0.12
0.32
0.08—1.33
Diabetes mellitus
0.71
1.81
0.35–9.24
Initial serum sodium (C155 mmol/L)
0.28
0.27
0.03–2.92
CNS symptomatology
0.61
1.56
0.29–8.38
Initial serum creatinine (C1.5 mg/dL)
0.56
1.54
0.37–6.45
mortality rate of 60 % was not different from that of patients whose serum sodium was corrected (p = 0.52). Of the 33 patients studied, 15 received a consultation by a nephrologist and of these, 5 survived and 10 died. Of the 18 patients without nephrology consultation, outcomes were not different, 11 lived and 7 died (p = 0.12). Thirteen of the 15 nephrology consultations obtained were within 1 day in regard to the recognition of the patients’ hypernatremic status. Eight of these patients died (62 %). In the 2 patients in which a nephrology consult was obtained[24 h after the recognition of hypernatremia, both died. Analysis of patient data also revealed that presence of diabetes mellitus, initial serum sodium concentration C155 mmol/ L, and CNS symptomatology did not influence mortality. Multivariable logistic regression analysis of variables versus mortality is depicted in Table 3. Thirty-one (94 %) patients received various IVF and/or oral fluids, including 0.9 % normal saline (NS), 0.45 % NS, D5-0.45 % NS, 0.3 % NS, D5, lactated ringers (LR), and oral water per os or via tube. The choice of fluid appeared appropriate to the clinical status with hypovolemic patients receiving normal saline, 0.45 % NS for uncertain volume status and water for those patients clearly only water deficient (Fig. 2). The mean length of stay for all patients was 25.0 ± 23.9 days and no different for those who expired versus those who remained alive (25.2 ± 21.2 vs. 24.8 ± 25.9 days, p = 0.96). Moreover, the mean length of stay was no different if a nephrology consultation was obtained (21.6 ± 20.8 vs. 20.1 ± 21.8 days, p = 0.84). Of the patients who presented from home (n = 20), 12 died, 5 returned to home, and 3 were sent to an advanced care facility. Of the patients who presented from an advanced care facility (n = 12), 4 died, zero went home, 5 returned to an advanced care facility, and 3 were sent to
123
Int Urol Nephrol
Fig. 2 Type of fluid selected to correct hypernatremia
Fig. 3 Discharge setting. Fifty-two percent of patients expired
hospice. The only patient who was transferred from an outside hospital died. Figure 3 summarizes the discharge settings of our patients.
Discussion In this retrospective study analyzing outcomes of hospitalized patients with hypernatremia, the mortality rate was 52 %, which is consistent with prior studies, with approximately half of the survivors being discharged to a facility with specialized services indicating the fragile nature of these patients. Hypernatremia was associated with advanced age, and the mean length of stay did not appear to be different for survivors compared to nonsurvivors. Cardiac-related disease, pulmonary-related disease, and sepsis comprised the majority of principle medical diagnoses related to hypernatremia. The mortality rate was found to be high despite prompt rate of correction, involvement of a medical specialist, and intensive care monitoring. Patients were more likely to expire whether they were admitted to or transferred to the ICU compared to receiving all of their care in a telemetry unit or in a general medical ward. The serum sodium concentration is usually maintained within a narrow range of 138–142 mmol/L despite great
123
variation in water intake among individuals. This is accomplished by regulating the balance of water through responses to osmoreceptors and baroreceptors that regulate the secretion of antidiuretic hormone (ADH) from the hypothalamus [6]. The fact that in our study, hypernatremia was associated with an elderly population is not surprising considering the physiologic changes associated with aging which predispose to hypernatremia as well as certain risk factors that increase their susceptibility to the condition. With advancing age, the ability to concentrate the urine diminishes and osmotic stimulation of thirst also diminishes, even though the mechanisms associated with the release of ADH are maintained [7–9]. It is speculated that the reduced ability to maximally concentrate the urine with aging may be related to decreased activity of many of the key transport proteins (aquaporin (AQP) 2, serine-256— phosphorylated AQP2, AQP3, NKCC2/BSC1, urea transporter genes UT-A1 and UT-B, and the V2-receptor) that contribute to medullary hypertonicity and water transport [10, 11]. Our study suggested that those patients managed in the ICU had a higher mortality rate (83 %) than those in nonintensive care areas, despite similar mean serum sodium levels. An explanation for this finding may be that such patients exhibited medical conditions which were more severe compared to those who were felt to be manageable in a non-ICU setting. In a study performed by Hoorn et al. [12] investigating the risk factors and mechanisms of hypernatremia in critically ill patients, they noted that the patients with ICU acquired hypernatremia were generally sicker on admission to the ICU, which was reflected by a higher Acute Physiology And Chronic Health Evaluation II (APACHE II) score and by a lower Glasgow Coma Score (GCS). The forty-five percent mortality observed in our study in patients who had hypernatremia corrected in 72 h or less is in contrast to earlier data. In a study by Alshayeb et al. evaluating the effect of the rate of correction of severe hypernatremia (defined as serum sodium C155 mEq/L) on the mortality of hospitalized patients, they found that patients whose hypernatremia was corrected within 72 h had a lower mortality (10 vs. 39 %, p = 0.0048) and were more likely to be managed in the ICU (89 vs. 64 %, p = 0.0096). This was observed in 107 patients who survived beyond 72 h [4]. Our study indicated that there was no benefit in regard to mortality for patients with hypernatremia corrected within 72 h. Factors that may influence the effect of prompt rate of correction may relate to a beneficial effect on more acutely hypernatremic patients as well as the severity of the patient’s condition. Indeed, our data revealed that survival was not improved whether or not the serum sodium was corrected. The diseases associated with hypernatremia in our patients included conditions
Int Urol Nephrol
with multi-organ failure in the setting of cardiac disease, advanced pneumonia, and sepsis indicating that disease severity was the major influence regarding outcome. Since hypernatremia is not a uniform disorder, treatment is often a challenging task. Current recommendations for the treatment for hypernatremia in elderly patients include calculation of the free water deficit [0.5 9 weight in kg] 9 [1 - (140/serum sodium)] and the correction of half the deficit in the first 24 h [13, 14]. Since ongoing losses of hypotonic fluids (i.e., gastrointestinal loss, pyrexia) may contribute, these losses should also be added to the total water deficit calculation. If there is concern that an osmotic diuresis is occurring, then measurements of plasma sodium, urine volume, and urine electrolyte concentrations (sodium and potassium) are used for the determination of electrolytefree water clearance (CeH2O = VU 9 (1 - ([Na]U ? [K]U)/[Na]S)) so that ongoing free water loss can also be added [15]. If the dysnatremia is acute over a few hours (e.g., iatrogenic sodium load), then it is appropriate to correct at a rate of 1–2 mmol/L/h until symptoms subside. If the change in sodium concentration occurred over a longer or unknown duration, the correction rate can be 0.5 mmol/L/h with a target of less than 12 mmol/L in a 24-h period to avoid cerebral edema and neurotoxicity [1]. Cerebral edema may actually be more of a concern in young children because of small skull volume [16, 17]. A hypotonic solution, such as pure water, D5, 0.3 % NS, or 0.45 % NS, should be used to decrease serum sodium. The preferred route of water administration is oral, per mouth, or through a feeding tube if necessary. If parenteral D5 is used, attention to correction of hyperglycemia is important. Normal saline is an appropriate initial therapy option if the patient is hypovolemic and must be fluid resuscitated [13]. Also, diuretics may be needed if intravascular volume excess develops or to assist with the excretion of hypertonic fluid. In summary, it is essential to diagnose the condition causing the hypernatremia along with assessing the patient’s volume status, estimating the water requirements, selecting a proper fluid type with an appropriate rate of correction, and monitoring the patient and laboratory values closely during therapy [18]. Our study differed from prior studies analyzing outcomes in patients with hypernatremia in several ways. First, we investigated certain variables in hospitalized hypernatremic geriatric patients which have not been previously evaluated. These included whether or not a nephrologist was involved in the case and the origin of the patient prior to admission. An interesting finding of our study was that the mortality rate observed in patients receiving a nephrology consultation was substantial (67 %; 5 alive, 10 died) compared to those who did not receive one
(39 %; 11 alive, 7 died) (p = 0.12). It seemed, therefore, that nephrology consultation did not appreciably influence patient outcomes. Review of patient data revealed that the timing of nephrology consultation often occurred early in the patient’s hospital course, but this also did not appear to affect the results noted. Secondly, our study quantitated the use of diagnostic urine laboratory measurements in our patients. Less than 25 % of patients had any urine studies performed which we feel would have helped in the correct evaluation of urinary concentrating ability and volume status [13]. Patients with low urine sodium levels and a concentrated urine could be classified to be volume depleted requiring volume resuscitation, while those with urine osmolalities suggesting a failure of urinary concentration would be those who required primarily water replacement. The observations regarding the limited use of plasma and urine laboratory studies are similar to those reported by Huda et al. [19] in a report describing hyponatremic patients and may represent a general lack of appreciation of the utility of these studies by clinicians. Our results indicate that hypernatremia continues to be associated with a high mortality rate. Prevention of hypernatremia may be more important than some of the therapeutic variables cited above. Our study revealed that approximately 40 % of patients presented from a facility other than home, most likely because the patient was unable to perform self-care. Since the geriatric population is already predisposed to the development of hypernatremia due to physiologic changes related to aging, certain measures could and should be taken for preventing hypernatremia, especially in those individuals that require supervision and an extra level of care (i.e., individuals in nursing homes). One such measure could be to obtain blood chemistries at more frequent intervals, particularly in patients receiving tube feeding or with limited mobility. This would be performed to not only assess the serum sodium level but also assess the serum urea and serum glucose levels since an osmotic diuresis may ensue from high serum concentration of these substances, thereby contributing to hypernatremia [20]. Another measure that could be employed would be the recording of input/output data more frequently with special attention to evaluation of fluid, especially water, intake. In conclusion, hypernatremia is associated with a poor prognosis, and outcomes are still disappointing despite appropriate rates of correction, intensive monitoring, and the involvement of a medical specialist. Implementation of strategies directed at the avoidance of hypernatremia, as well as vigilance regarding the development of new or the worsening of concomitant disorders, may provide patient benefit.
123
Int Urol Nephrol Conflict of interest The authors of this manuscript declare that they have no conflict of interest.
References 1. Adrogue´ HJ, Madias NE (2000) Hypernatremia. N Engl J Med 342(20):1493–1499 2. Lindner G, Funk GC, Schwarz C, Kneidinger N, Kaider A, Schneeweiss B, Druml W (2007) Hypernatremia in the critically ill is an independent risk factor for mortality. Am J Kidney Dis 50:952–957. doi:10.1053/j.ajkd.2007.08.016 3. Palevsky PM, Bhagrath R, Greenberg A (1996) Hypernatremia in hospitalized patients. Ann Intern Med 124:197–203. doi:10.7326/ 0003-4819-124-2-199601150-00002 4. Alshayeb HM, Showkat A, Babar F, Mangold T, Wall BM (2011) Severe hypernatremia correction rate and mortality in hospitalized patients. Am J Med Sci 341:356–360. doi:10.1097/MAJ. 0b013e31820a3a90 5. Mandal AK, Saklayen MG, Hillman NM, Markert RJ (1997) Predictive factors for high mortality in hypernatremia patients. Am J Emerg Med 15:130–132. doi:10.1016/S07356757(97)90082-6 6. Kumar S, Berl T (1998) Sodium. Lancet 352:220–228. doi:10. 1016/S0140673697121699 7. Rowe JW, Shock NW, DeFronzo R (1976) The influence of age on the renal response to water deprivation in man. Nephron 17:270–278. doi:10.1159/000180731 8. Kenney WL, Chiu P (2001) Influence of age on thirst and fluid intake. Med Sci Sports Exerc 33:1524–1532. doi:10.1097/ 00005768-200109000-00016 9. Rose BD, Post TW (2001) Hyperosmolal states–hypernatremia. In: Clinical physiology of acid-base and electrolyte disorders, 5th edn. McGraw-Hill, New York, pp 746–793
123
10. Sands JM (2009) Urinary concentration and dilution in the aging kidney. Semin Nephrol 29:579–586. doi:10.1016/j.semnephrol. 2009.03.008 11. Zhou XJ, Saxena R, Liu Z, Vaziri ND, Silva FG (2008) Renal senescence in 2008: progress and challenges. Int Urol Nephrol 40:823–839. doi:10.1007/s11255-008-9405-0 12. Hoorn EJ, Betjes MG, Weigel J, Zietse R (2008) Hypernatraemia in critically ill patients: too little water and too much salt. Nephrol Dial Transplant 23:1562–1568. doi:10.1093/ndt/gfm831 13. Michelis MF (2009) Chapter 16: disorders of serum sodium concentration in the elderly patient. In: Geriatric nephrology curriculum, American Society of Nephrology pp 1–6 14. Adrogue HJ, Madias NE (1997) Aiding fluid prescription for the dysnatremias. Intensive Care Med 23:309–316. doi:10.1007/ s001340050333 15. Popli S, Tzamaloukas AH, Ing TS (2013) Osmotic diuresisinduced hypernatremia: better explained by solute-free water clearance or electrolyte-free water clearance? Int Urol Nephrol (Epub January 2013). doi 10.1007/s11255-012-0353-3 16. Ayus JC, Brennan S (1995) Hypernatremia. In: Arieff AI, DeFronzo RA (eds) Fluid, electrolyte, and acid-base disorders, 2nd edn. Churchill Livingstone, New York, pp 304–317 17. Xenos C, Sgouros S, Natarajan K (2002) Ventricular volume change in childhood. J Neurosurg 97:584–590 18. Al-Absi A, Gosmanova EO, Wall BM (2012) A clinical approach to the treatment of chronic hypernatremia. Am J Kidney Dis 60:1032–1038. doi:10.1053/j.ajkd.2012.06.025 19. Huda MB, Boyd A, Skagen K, Wile D, Van Heyningen C, Watson I, Wong S, Gill G (2006) Investigation and management of severe hyponatraemia in a hospital setting. Postgrad Med J 82:216–219. doi:10.1136/pmj.2005.036947 20. Lindner G, Schwarz C, Funk GC (2012) Osmotic diuresis due to urea as the cause of hypernatraemia in critically ill patients. Nephrol Dial Transplant 27:962–967. doi:10.1093/ndt/gfr428
