- Email: [email protected]
Sodium Correction Practice and Clinical Outcomes in Profound Hyponatremia
ORIGINAL ARTICLE
Sodium Correction Practice and Clinical Outcomes in Profound Hyponatremia Pierce Geoghegan, MB BCh; Andrew M. Harrison, BS; Charat Thongprayoon, MD; Rahul Kashyap, MD; Adil Ahmed, MBBS; Yue Dong, MD; Alejandro A. Rabinstein, MD; Kianoush B. Kashani, MD; and Ognjen Gajic, MD, MSc Abstract Objectives: To assess the epidemiology of nonoptimal hyponatremia correction and to identify associated morbidity and in-hospital mortality. Patients and Methods: An electronic medical record search identified all patients admitted with profound hyponatremia (sodium <120 mmol/L) from January 1, 2008, through December 31, 2012. Patients were classified as having optimally or nonoptimally corrected hyponatremia at 24 hours after admission. Optimal correction was defined as sodium correction in 24 hours of 6 through 10 mmol/L. We investigated the association between sodium correction and demographic and outcome variables, including occurrence of osmotic demyelination syndrome (ODS). Baseline characteristics by correction outcome categories were compared using the Kruskal-Wallis test for continuous variables and the c2 test for categorical variables. Odds ratios for in-hospital mortality between groups were assessed using logistic regression. Adjusted differences in hospital length of stay (LOS) and intensive care unit (ICU) LOS were assessed using the Dunnett 2-tailed t test. Results: A total of 412 patients satisfied inclusion criteria of whom 174 (42.2%) were admitted to the ICU. A total of 211 (51.2%) had optimal correction of their hyponatremia at 24 hours, 87 (21.1%) had undercorrected hyponatremia, and 114 (27.9%) had overcorrected hyponatremia. Both patient factors and treatment factors were associated with nonoptimal correction. There was a single case of ODS. Overcorrection was not associated with in-hospital mortality or ICU LOS. When adjusted for patient factors, undercorrection of profound hyponatremia was associated with an increase in hospital LOS (9.3 days; 95% CI, 1.9-16.7 days). Conclusion: Nonoptimal correction of profound hyponatremia is common. Fortunately, nonoptimal correction is associated with serious morbidity only infrequently. ª 2015 Mayo Foundation for Medical Education and Research
For editorial comment, see page 1320 From the Multidisciplinary Epidemiology and Translational Research in Intensive Care (P.G., A.M.H., C.T., R.K., A.A., Y.D., A.A.R., K.B.K., O.G.), Medical Scientist Training Program (A.M.H.), Department of Anesthesiology (R.K.), Division of Pulmonary and Critical Care Affiliations continued at the end of this article.
1348
H
yponatremia is a common electrolyte abnormality in hospitalized patients,1 and dysnatremia is an independent risk factor for hospital mortality.2 Profound hyponatremia (serum sodium <120 mmol/L) can lead to cerebral edema, and correction of the hyponatremia is necessary to prevent or treat associated neurologic complications or even death.3 Cerebral edema is a particular risk when the hyponatremia is acute (48 hours) because the brain has not yet had sufficient time to make adaptations to the hypotonic environment.4,5 In chronic hyponatremia (>48 hours), the brain has usually made adaptations to the hypotonic environment, and overly rapid
n
Mayo Clin Proc. 2015;90(10):1348-1355
correction of profound hyponatremia is associated with osmotic demyelination syndrome (ODS), which may itself cause substantial morbidity or even mortality.6 Thus, the duration of hyponatremia, if known accurately, can determine whether the primary risk to the patient is from cerebral edema or ODS. Experts have disagreed on exactly what constitutes optimal correction.7 Deciding what constitutes optimal correction is a particular challenge in patients who have hyponatremia at admission (rather than develop it during hospitalization) because the duration of hyponatremia is usually not known.8 Therefore, it can be argued that sodium correction in these patients
Mayo Clin Proc. n October 2015;90(10):1348-1355 n http://dx.doi.org/10.1016/j.mayocp.2015.07.014 www.mayoclinicproceedings.org n ª 2015 Mayo Foundation for Medical Education and Research
SODIUM CORRECTION AND OUTCOMES IN HYPONATREMIA
must account for the risk of cerebral edema (if acute) or the risk of ODS (if chronic). There is good evidence to suggest that raising the serum sodium level by a modest 4 to 6 mmol/L is more than sufficient to treat clinically significant neurologic symptoms due to profound hyponatremia.9 Data from case reports and case series suggest that there is probably a continuum of risk of ODS in chronic hyponatremia that increases when correction is faster than 8 mmol/L per 24 hours and is highest if correction rates exceed 12 to 14 mmol/L per 24 hours.10-12 The most recent guidelines recommend a ceiling of correction of 10 mmol per 24 hours or less when there is a risk of ODS and an initial correction target of 5 mmol per 24 hours to treat any symptoms of cerebral edema.8 Treatment response is unpredictable in profound hyponatremia.13,14 Nevertheless, the epidemiology of nonoptimal correction of profound hyponatremia is not well described. In particular, the risk factors for nonoptimal correction and its implications for patient outcomes have been poorly delineated. We hypothesized that nonoptimal correction of profound hyponatremia would be common and associated with increased morbidity and in-hospital mortality. In the current study, our objective was to assess the epidemiology of nonoptimal correction in patients at Mayo Clinic’s campus in Rochester, Minnesota, and to attempt to identify any associated morbidity (including ODS) and mortality. PATIENTS AND METHODS Study Design and Setting Our study was a retrospective analysis of our longstanding electronic medical record (EMR) system at the Rochester campus of Mayo Clinic. The local institutional review board reviewed and approved the project. We performed a search of the EMR to identify all patients admitted to our institution from January 1, 2008, through December 31, 2012, with a serum sodium level less than 120 mmol/L after correction for serum glucose using a well-described correction factor.15 We excluded patients younger than 18 years, patients who did not provide research authorization, those who were diagnosed as having pseudohyponatremia, and, finally, Mayo Clin Proc. n October 2015;90(10):1348-1355 www.mayoclinicproceedings.org
n
those who died in the emergency department or discharged against medical advice in the first 24 hours of treatment. When patients were admitted with profound hyponatremia more than once during the study period, only the first admission was considered in the final analysis. Study Participants and Data Collection For each patient included in the study, we further identified all serum sodium measurements undertaken in the first 72 hours of admission. Taking the time of the initial sodium check (Na0) identifying profound hyponatremia as time zero (T0), we estimated serum sodium levels at the 24-hour mark (Na24). To do this, we used the formula: Na24 ¼ Naa þ [(Nab Naa) (24 Ta)/Tb Ta)], where Naa indicates the closest serum sodium measurement before the 24-hour mark; Nab, the closest serum sodium measurement after the 24-hour mark; Ta, the time at which measurement Naa was taken; and Tb, the time at which measurement Nab was taken. The degree of correction at 24 hours (DNa24) was calculated as Na24 e Na0. Patients were judged to be optimally corrected if their DNa24 levels were 6 through 10 mmol/L. All other patients were classified as having nonoptimally corrected hyponatremia. Patients whose hyponatremia was nonoptimally corrected were further divided into those with undercorrected hyponatremia (DNa24 5 mmol/L) and overcorrected hyponatremia (DNa24 >10 mmol/L). To establish whether individual patients had experienced clinically or radiologically confirmed ODS, we performed an electronic search of the EMR to identify terms in the treating or consulting physicians notes that suggested suspicion or confirmation of a diagnosis of ODS during the admission with profound hyponatremia or up to 6 months thereafter. These terms included central pontine myelinolysis, extrapontine myelinolysis, and osmotic demyelination syndrome, as well as combinations of these. Individual medical records that contained these terms were then manually reviewed by a clinician to determine whether a diagnosis of ODS had indeed been confirmed clinically or radiologically. The records of a random sample of 10% of all patients whose EMRs did not contain any of these terms were also manually
http://dx.doi.org/10.1016/j.mayocp.2015.07.014
1349
MAYO CLINIC PROCEEDINGS
All patients admitted between January 1, 2008, through December 31, 2012, were considered for eligibility
540 Admission serum sodium <120 mmol/L
128 Excluded : • 48 Age <18 years at admission • 50 Absence of research authorization • 26 Pseudohyponatremia • 4 Discharged against medical advice or died within 24 hours 412 Met all inclusion criteria
Classified according to estimated serum sodium at 24 hours
87 Undercorrection (≤5 mmol/L at 24 hours)
211 Optimal correction (6-10 mmol/L at 24 hours)
114 Overcorrection (>10 mmol/L at 24 hours)
FIGURE. Patient selection process and subsequent categorization according to degree of correction of profound hyponatremia 24 hours after the initial sodium measurement.
searched to ensure that cases of ODS were not being missed. We additionally cross-referenced our search with a previously published cohort of ODS cases from our institution to ensure that no cases of ODS in our cohort had been missed.16 A similar strategy was used to identify instances where clinicians had identified one of a (nonexhaustive) list of common causes of hyponatremia (eg, hypovolemia, thiazide diuretic use) on initial assessment of each patient and to identify the administration of therapies, such as hypertonic saline or desmopressin. The EMR was also used to identify other relevant data, including demographic data and clinical outcome measures, using previously validated methods and the Charlson definitions for individual comorbidities, including congestive heart failure, chronic kidney disease, and cirrhosis.17,18 Primary outcomes included the epidemiology of nonoptimal correction, the incidence of clinically or radiologically confirmed ODS, in-hospital mortality, and length of stay (LOS). Secondary outcomes were the association 1350
Mayo Clin Proc.
n
between patient and treatment factors and risk of nonoptimal correction. Statistical Analyses Patient demographic characteristics were summarized using median and interquartile range (IQR) for continuous variables and number and percentage for categorical variables. The baseline characteristics by sodium outcome categories were compared using the Kruskal-Wallis test for continuous variables and the c2 test for categorical variables. The adjusted and unadjusted odds ratios for in-hospital mortality in patients with undercorrected, optimally corrected, and overcorrected hyponatremia were assessed using a logistic regression with the in-hospital mortality for the patients with optimally corrected hyponatremia being used as the referent group. The independent variables used in this multivariable logistic regression were hypertonic saline therapy, age, Charlson comorbidity score, congestive heart failure, cirrhosis or chronic liver disease, moderate or severe chronic kidney disease, and baseline sodium level. The unadjusted
October 2015;90(10):1348-1355
n
http://dx.doi.org/10.1016/j.mayocp.2015.07.014 www.mayoclinicproceedings.org
Characteristic Age, y No. of participants Female Male Total BMI Charlson comorbidity score Admission sodium, mmol/L Admission serum osmolarity, mOsmol/kg
n
http://dx.doi.org/10.1016/j.mayocp.2015.07.014
Admission urine osmolarity, mOsmol/kg Admission urine sodium, mmol/L Patients with admission urine sodium <20 mmol/L Patients with admission potassium <3 mEq/L Seizures Congestive heart failure Diagnosis of cirrhosis or chronic liver disease Moderate to severe chronic kidney disease Acute kidney injury within 24 hours Diagnosis of alcoholism at admission Taking a thiazide diuretic before admission Taking a SSRI before admission Diagnosis of hypovolemia at admission Diagnosis of pneumonia at admission SIADH in admission differential diagnosis Admitted to ICU 3-Hour APACHE III score Total fluid in, L Total fluid out, L Hypertonic saline therapy Desmopressin therapy Vasopressin receptor antagonist therapy Number of sodium checks DNR/DNI order in place Nephrology consultation Received dialysis
Undercorrection (n¼87)
Optimal correction (n¼211)
Overcorrection (n¼114)
Total (N¼412)
P value
69 (55-79) (n¼86)
71 (60-81) (n¼210)
63 (51-78) (n¼114)
69 (56-80) (n¼410)
.009
52 (59.8) 35 (40.2) 87 28 (21-33) (n¼87) 6 (4-8) (n¼87) 117 (115-119) (n¼87) 248 (243-254) (n¼58) 58 363 (291-466) (n¼71) 39 (23-66) (n¼26) 5 (19.2) (n¼26) 3 (3.5) (n¼86) 1 (1.1) (n¼87) 41 (47.1) (n¼87) 16 (18.4) (n¼87) 14 (16.1) (n¼87) 11 (12.6) (n¼87) 12 (13.8) (n¼87) 16 (18.4) (n¼87) 19 (21.8) (n¼87) 25 (28.7) (n¼87) 19 (21.8) (n¼87) 20 (23.0) (n¼87) 27 (31.0) (n¼87) 56 (37-63) (n¼27) 1.1 (0.7-2.5) (n¼39) 1.3 (0.7-2.2) (n¼39) 4 (4.6) (n¼87) 0 (0.0) (n¼87) 1 (1.1) (n¼87) 3 (2-5) (n¼87) 15 (17.2) (n¼87) 17 (19.5) (n¼87) 0 (0) (n¼87)
123 (58.3) 88 (41.7) 211 25 (22-30) (n¼206) 5 (3-8) (n¼211) 117 (114-119) (n¼211) 250 (241-258) (n¼138) 138 338 (266-429) (n¼180) 47 (21-80) (n¼67) 13 (19.4) (n¼67) 24 (11.4) (n¼211) 10 (4.7) (n¼211) 76 (36.0) (n¼211) 10 (4.7) (n¼211) 31 (14.7) (n¼211) 44 (20.9) (n¼211) 38 (18.0) (n¼211) 46 (21.8) (n¼211) 54 (25.6) (n¼211) 79 (37.4) (n¼211) 56 (26.5) (n¼211) 55 (26.0) (n¼211) 91 (43.1) (n¼211) 52 (42-68) (n¼91) 3.0 (1.4-4.2) (n¼105) 2.0 (1.2-3.0) (n¼105) 11 (5.2) (n¼211) 2 (0.9) (n¼211) 1 (0.5) (n¼211) 5 (3-7) (n¼211) 42 (19.9) (n¼211) 40 (19.0) (n¼211) 5 (2.3) (n¼211)
65 (57.0) 49 (43.0) 114 24 (22-29) (n¼112) 5 (3-8) (n¼114) 115 (112-117) (n¼114) 249 (239-258) (n¼74) 74 304 (199-384) (n¼91) 39 (19-65) (n¼29) 8 (27.6) (n¼29) 21 (18.4) (n¼114) 7 (6.1) (n¼114) 20 (17.5) (n¼114) 3 (2.6) (n¼114) 7 (6.1) (n¼114) 20 (17.5) (n¼114) 17 (14.9) (n¼114) 29 (25.4) (n¼114) 26 (22.8) (n¼114) 40 (35.1) (n¼114) 25 (21.9) (n¼114) 29 (25.4) (n¼114) 56 (49.1) (n¼114) 55 (38-72) (n¼56) 4.4 (2.4-8.3) (n¼72) 3.9 (2.3-6.5) (n¼72) 14 (12.3) (n¼114) 2 (1.8) (n¼114) 1 (0.9) (n¼114) 6 (4-8) (n¼114) 10 (8.8) (n¼114) 13 (11.4) (n¼114) 2 (1.7) (n¼114)
240 (58.3) 172 (41.7) 412 25 (22-30) (n¼405) 5 (3-7) (n¼412) 116 (114-118) (n¼412) 249 (242-257) (n¼270) 270 332 (256-427) (n¼342) 43 (21-73) (n¼122) 26 (21.3) (n¼122) 48 (11.7) (n¼411) 18 (4.4) (n¼412) 137 (33.3) (n¼412) 29 (7.0) (n¼412) 52 (12.6) (n¼412) 75 (18.2) (n¼412) 67 (16.3) (n¼412) 91 (22.1) (n¼412) 99 (24.0) (n¼412) 144 (34.9) (n¼412) 100 (24.2) (n¼412) 104 (25.2) (n¼412) 174 (42.2) (n¼412) 53 (41-68) (n¼174) 2.8 (1.3-5.1) (n¼216) 2.2 (1.3-3.9) (n¼216) 29 (7.0) (n¼412) 4 (1.0) (n¼412) 3 (0.7) (n¼412) 5 (3-7) (n¼412) 67 (16.2) (n¼412) 70 (17.0) (n¼412) 7 (1.7) (n¼412)
.93
.13 .002 <.001 .66 .003 .76 .64 .005 .21 <.001 <.001 .04 .24 .60 .49 .74 .36 .55 .86 .03 .82 <.001 <.001 .04 .45 .80 <.001 .03 .17 .35
APACHE ¼ Acute Physiology and Chronic Health Evaluation; BMI ¼ body mass index; DNR/DNI ¼ do not resuscitate/do not intubate; ICU ¼ intensive care unit; SIADH ¼ syndrome of inappropriate antidiuretic hormone secretion; SSRI ¼ selective serotonin reuptake inhibitor. b Data are presented as median (interquartile range) for continuous variables and number (percentage) for categorical variables. a
SODIUM CORRECTION AND OUTCOMES IN HYPONATREMIA
Mayo Clin Proc. n October 2015;90(10):1348-1355 www.mayoclinicproceedings.org
TABLE 1. Association of Patient and Treatment Factors to Sodium Correction Outcome at 24 Hoursa,b
1351
MAYO CLINIC PROCEEDINGS
TABLE 2. Primary and Secondary Outcome Measuresa,b Outcome measure
Undercorrection (n¼87)
Osmotic demyelination syndrome 0 (0) Survival to discharge 78 (88) Odds ratio (95% CI) of death at discharge Unadjusted 2.1 (0.8-5.3) Adjustedc 2.2 (0.8-5.6) In-hospital length of stay, d 5 (3-8) ICU length of stay, d 2 (1-4) Between group differences in hospital length of stay, d Unadjusted 9 (2-17) Adjustedc 9 (2-17) Between-group difference in ICU length of stay, d Unadjusted 1 (1 to 2) Adjusted 1 (1 to 2)
Optimal correction (n¼211) 0 (0) 200 (95) 1.0 [Reference] 1.0 [Reference] 5 (3-8) 2 (1-3)
Overcorrection (n¼114)
Total P (N¼412) value
1 (1) 108 (95)
1 (0) 386 (94)
1.0 1.0 4 2
(0.4-2.8) (0.4-3.1) (2-8) (1-3)
. . 5 (3-8) 2 (1-3)
.27 .22 .23 .24 .10 .98
0 [Reference] 0 [Reference]
1 (6 to 7) 1 (6 to 7)
. .
.02 .04
0 [Reference] 0 [Reference]
1 (1 to 2) 1 (1 to 2)
. .
.4 .58
ICU ¼ intensive care unit. Data are reported as median (interquartile range) or median (range) for continuous variables and number (percentage) for categorical variables. Between group differences in hospital and ICU length of stay are reported as difference in days (95% confidence interval for difference). c Adjusted for use of hypertonic saline, age at admission, Charlson comorbidity score, congestive cardiac failure, cirrhosis or chronic liver disease, moderate to severe chronic kidney disease, and corrected admission serum sodium level. a
b
and adjusted difference in the hospital LOS and intensive care unit (ICU) LOS was assessed using Dunnett 2-tailed t test using the LOS for patients with optimally corrected hyponatremia as the referent group. An adjusted analysis was used to reduce confounding due to being a retrospective study. The Wilson score method was used to calculate a 95% CI for the incidence of ODS in the overcorrection group. P < .05 was considered statistically significant. Statistical analyses were performed using JMP statistical software (SAS Institute Inc). RESULTS The Figure shows how patients were identified for inclusion in the study, as well as their subsequent grouping according to degree of sodium correction at 24 hours. A total of 412 patients were included; 211 patients (51.2%) had optimally corrected hyponatremia at 24 hours, 87 patients (21.1%) had undercorrected hyponatremia, and 114 (27.7%) had overcorrected hyponatremia according to our predefined criteria. The closest sodium measurements to the 24-hour mark (used to estimate Na24) were a median of 3.6 hours before (IQR hyponatremia, 1.6-7.5) and 3.7 hours after (IQR, 1.5-9.0) the 24-hour mark. Table 1 lists the patient characteristics and treatment factors and their association with the adequacy of sodium correction outcome at 24 1352
Mayo Clin Proc.
n
hours. Overall, the median age in the study population was 69 years, and 240 patients (58.2%) were female. Median serum sodium on admission after correcting for blood glucose level was 116 mmol/L. In total, 174 patients (42.2%) were admitted to an ICU. Patients were more likely to experience overcorrection if they were younger, had lower admission serum sodium values, experienced seizures, required ICU admission, or received hypertonic saline (Table 1). Of importance, patients who were admitted to the ICU were more likely to have experienced seizures than those treated on a ward (9.2% vs 0.8%, P<.001) and were also more likely to receive hypertonic saline (10.3% vs 4.6%, P¼.02). This could mean that more symptomatic patients were admitted to the ICU and required more aggressive correction. Overcorrection was also associated with larger fluid input and with larger fluid output in the first 24 hours of treatment and with lower urine osmolality at first measurement. Patients were more likely to experience undercorrection if they had higher Charlson comorbidity scores and specific comorbidities, including congestive heart failure, cirrhosis or chronic liver disease, or moderate to severe chronic kidney disease. Table 2 gives a summary of the major primary outcomes, again grouped according to degree of 24-hour sodium correction. A single
October 2015;90(10):1348-1355
n
http://dx.doi.org/10.1016/j.mayocp.2015.07.014 www.mayoclinicproceedings.org
SODIUM CORRECTION AND OUTCOMES IN HYPONATREMIA
case of ODS was observed in the overcorrected group, corresponding to an incidence of approximately 1% (95% CI, 0%-5%) in those who were overcorrected in the first 24 hours of treatment. This patient was a 40-year-old woman with a diagnosis of psychogenic polydipsia who presented with a serum sodium level of 97 mmol/L and experienced a sodium correction of 21 mmol/L in 24 hours. The patient did not have a diagnosis of alcoholism and did not experience any hypoxia in the first 24 hours of admission. In an adjusted analysis, undercorrection at 24 hours was associated with a statistically significant increase in hospital LOS of 9 days (95% CI, 2-17 days). Nonoptimal correction was not associated with a statistically significant change in ICU LOS or in hospital mortality. DISCUSSION Our results indicate that nonoptimal correction of hyponatremia was very common in patients admitted with profound hyponatremia, occurring in nearly half of the patients in this cohort. Although it was frequent, overcorrection was associated with only a single case of ODS (corresponding to an incidence of approximately 1% in those who were overcorrected) and was not associated with changes in hospital mortality or LOS. Hospital LOS was increased in patients who met our definition of undercorrection 24 hours after admission. The high incidence of overcorrection is perhaps not surprising when one considers that response to treatment in profound hyponatremia is known to be unpredictable.14 In addition, the incidence of overcorrection in our cohort is similar to that which was observed in a similar cohort, when adjusting for a different definition of overcorrection.19 This finding suggests that this is a widespread problem rather than one limited to our hospital. That said, we admit that other institutions have described lower rates of overcorrection, albeit in selected cohorts.20 Overcorrection was more common in patients admitted to the ICU. This may be because patients with symptoms of cerebral edema (eg, seizures) were more likely to be admitted to the ICU and subsequently receive aggressive correction therapy, including the use of hypertonic saline. Overcorrection was less common in patients who had higher Charlson comorbidity scores or who had active do-not-resuscitate or Mayo Clin Proc. n October 2015;90(10):1348-1355 www.mayoclinicproceedings.org
n
do-not-intubate orders. It is possible to speculate that this was because these patients were less likely to have aggressive treatment in general, including less aggressive correction of hyponatremia. Our observed incidence of ODS in those who had overcorrected hyponatremia was approximately 1% (95% CI, 0%-5%), much lower than the rate of 11% reported in the aforementioned study, which had a similar rate of overcorrection.19 The median sodium correction in the overcorrected group was 14 mmol/L (IQR, 12-16 mmol/L) so that it is possible that this is partly related to the fact that often the targets were only marginally missed in our cohort. Indeed, the only patient in our cohort who developed ODS had a rate of correction that was 2-fold faster than the maximum recommended. Clinicians should also bear in mind that the risk of ODS increases with faster rates of correction and greater degree of correction and is further modified by patient factors such as chronicity of hyponatremia, underlying nutritional status, history of alcoholism, liver transplantation, and presence of simultaneous hypoxia or sepsis.16,21,22 Some of these risk factors (eg, hypoxia) are modifiable, and some of the risk of ODS in our cohort may have been mitigated by good multidisciplinary care. Clinicians should remember that even if the risk of ODS is low, it can result in severe morbidity when it occurs. Aggressive correction is also ill-advised because an increase in the serum sodium of 4 to 6 mmol/L per 24 hours generally suffices to reverse any neurologic symptoms, and no benefit has been found with faster correction.23 Undercorrection was associated with a clinically and statistically significant increase in overall hospital LOS of 9.3 days (95% CI, 1.9-16.7 days). This was true even after adjusting for confounders that are traditionally thought to explain the association between increased morbidity and in-hospital mortality and hyponatremia, including age, diagnosis of congestive heart failure, liver disease, chronic kidney disease, or a diagnosis of malignant disease.24 We believe that the most plausible explanation for this association is that treating clinicians kept patients in the hospital longer to correct the hyponatremia, as opposed to the hyponatremia itself producing genuine morbidity. Nevertheless, it
http://dx.doi.org/10.1016/j.mayocp.2015.07.014
1353
MAYO CLINIC PROCEEDINGS
is possible that unrecognized or suboptimally treated cerebral edema produced genuine morbidity or that morbidity was produced through an alternate mechanism. Indeed, there is some evidence that early correction of profound hyponatremia (while maintaining correction rates lower than those associated with ODS) is associated with improved short-term outcomes, although this point has not been rigorously investigated.5 In our cohort, undercorrection at 24 hours was not associated with a statistically significant increase in the risk of in-hospital mortality, although it should be borne in mind that our study may not have been large enough to detect small differences in this end point. Our study has several limitations. First, the study design was retrospective. Second, it is a single-center study, and the findings would require confirmation in future multicenter studies. Third, our definitions of undercorrection and overcorrection are according to our interpretation of available evidence but are open to debate because experts have disagreed on what constitutes optimal correction. Fourth, because this study was observational, we had to estimate the serum sodium level at 24 hours because most patients did not have a serum sodium measurement at exactly this point in time. Fifth, sodium correction between 24 and 48 hours after admission was not analyzed and may influence the risk of ODS. Sixth, these definitions of nonoptimal correction are context specific (eg, the concept of overcorrection applies only where duration of hyponatremia is >48 hours). Similarly, the concept of undercorrection has been traditionally applied only to patients who are symptomatic (although this may not be a very practical distinction because there is increasing doubt that any of these patients are truly asymptomatic and patients with acute hyponatremia can present without clear symptoms but later develop severe symptoms within hours4,8). We considered our cohort as susceptible to the complications of both acute and chronic hyponatremia, although in reality each patient can only have been one or the other. However, we would argue that this approach is supported by recent guidelines that highlight that the duration of hyponatremia in these patients will generally be unknown.8 In addition, we did not identify 1354
Mayo Clin Proc.
n
which patients experienced symptoms that were ascribed to hyponatremia. CONCLUSION In the current study, we identified a high rate of nonoptimal sodium correction during the first 24 hours of treatment in a large unselected cohort of patients admitted to our institution with profound hyponatremia during a 5-five year period. Fortunately, this high rate of nonoptimal correction was not associated with increased in-hospital mortality and seemed to be associated with morbidity only infrequentlydODS occurred in only a single patient, corresponding to an incidence of approximately 1% in patients who had overcorrected hyponatremia. ACKNOWLEDGMENTS The authors gratefully acknowledge the entire Multidisciplinary Epidemiology and Translational Research in Intensive Care research group for their support, which made this project possible. The authors also thank Mayo Clinic’s Center for Translational Science Activities for their support. The contents of this publication are solely the responsibility of the authors and do not necessarily represent the official views of the National Institutes of Health. Affiliations (Continued from the first page of this article.): Medicine (Y.D., O.G.), Department of Neurology (A.A.R.), and Division of Nephrology and Hypertension (K.B.K.), Mayo Clinic, Rochester, MN; and Wichita Falls Family Practice Residency Program, North Central Texas Medical Foundation, Wichita Falls (A.A.). Grant Support: The work was supported by the Mayo Clinic’s Center for Translational Science Activities grant UL1 TR000135 from the National Center for Advancing Translational Sciences. Correspondence: Address to Pierce Geoghegan, MB, BCh, BAO, BA, Department of Anaesthesia and Critical Care, Rotunda Hospital, Dublin, Ireland ([email protected]).
REFERENCES 1. DeVita MV, Gardenswartz MH, Konecky A, Zabetakis PM. Incidence and etiology of hyponatremia in an intensive care unit. Clin Nephrol. 1990;34(4):163-166. 2. Darmon M, Diconne E, Souweine B, et al. Prognostic consequences of borderline dysnatremia: pay attention to minimal serum sodium change. Crit Care. 2013;17(1):R12. 3. Schrier RW, Bansal S. Diagnosis and management of hyponatremia in acute illness. Curr Opin Crit Care. 2008;14(6):627-634. 4. Arieff AI. Hyponatremia, convulsions, respiratory arrest, and permanent brain damage after elective surgery in healthy women. N Engl J Med. 1986;314(24):1529-1535.
October 2015;90(10):1348-1355
n
http://dx.doi.org/10.1016/j.mayocp.2015.07.014 www.mayoclinicproceedings.org
SODIUM CORRECTION AND OUTCOMES IN HYPONATREMIA
5. Nzerue CM, Baffoe-Bonnie H, You W, Falana B, Dai S. Predictors of outcome in hospitalized patients with severe hyponatremia. J Natl Med Assoc. 2003;95(5):335-343. 6. Karp BI, Laureno R. Pontine and extrapontine myelinolysis: a neurologic disorder following rapid correction of hyponatremia. Medicine (Baltimore). 1993;72(6):359-373. 7. Overgaard-Steensen C. Initial approach to the hyponatremic patient. Acta Anaesthesiol Scand. 2011;55(2):139-148. 8. Spasovski G, Vanholder R, Allolio B, et al. Clinical practice guideline on diagnosis and treatment of hyponatraemia. Nephrol Dial Transplant. 2014;29(suppl 2):i1-i39. 9. Sterns RH, Hix JK, Silver S. Treatment of hyponatremia. Curr Opin Nephrol Hypertens. 2010;19(5):493-498. 10. Adrogué HJ, Madias NE. Hyponatremia. N Engl J Med. 2000; 342(21):1581-1589. 11. Sterns RH, Nigwekar SU, Hix JK. The treatment of hyponatremia. Semin Nephrol. 2009;29(3):282-299. 12. Brunner JE, Redmond JM, Haggar AM, Kruger DF, Elias SB. Central pontine myelinolysis and pontine lesions after rapid correction of hyponatremia: a prospective magnetic resonance imaging study. Ann Neurol. 1990;27(1):61-66. 13. Vaidya C, Ho W, Freda BJ. Management of hyponatremia: providing treatment and avoiding harm. Cleve Clin J Med. 2010;77(10):715-726. 14. Nguyen MK. Quantitative approaches to the analysis and treatment of the dysnatremias. Semin Nephrol. 2009;29(3):216-226. 15. Hillier TA, Abbott RD, Barrett EJ. Hyponatremia: evaluating the correction factor for hyperglycemia. Am J Med. 1999;106(4): 399-403.
Mayo Clin Proc. n October 2015;90(10):1348-1355 www.mayoclinicproceedings.org
n
16. Graff-Radford J, Fugate JE, Kaufmann TJ, Mandrekar JN, Rabinstein AA. Clinical and radiologic correlations of central pontine myelinolysis syndrome. Mayo Clin Proc. 2011;86(11):1063-1067. 17. Ahmed A, Thongprayoon C, Pickering BW, et al. Towards prevention of acute syndromes: electronic identification of at-risk patients during hospital admission. Appl Clin Inform. 2014;5(1):58-72. 18. Singh B, Singh A, Ahmed A, et al. Derivation and validation of automated electronic search strategies to extract Charlson comorbidities from electronic medical records. Mayo Clin Proc. 2012;87(9):817-824. 19. Vu T, Wong R, Hamblin PS, Zajac J, Grossmann M. Patients presenting with severe hypotonic hyponatremia: etiological factors, assessment, and outcomes. Hosp Pract (1995). 2009;37(1):128-136. 20. Sood L, Sterns RH, Hix JK, Silver SM, Chen L. Hypertonic saline and desmopressin: a simple strategy for safe correction of severe hyponatremia. Am J Kidney Dis. 2013;61(4):571-578. 21. Martin RJ. Central pontine and extrapontine myelinolysis: the osmotic demyelination syndromes. J Neurol Neurosurg Psychiatry. 2004;75(suppl 3):iii22-iii28. 22. Singh TD, Fugate JE, Rabinstein AA. Central pontine and extrapontine myelinolysis: a systematic review. Eur J Neurol. 2014;21(12):1443-1450. 23. Verbalis JG, Goldsmith SR, Greenberg A, Schrier RW, Sterns RH. Hyponatremia treatment guidelines 2007: expert panel recommendations. Am J Med. 2007;120(11 suppl 1):S1-S21. 24. Chawla A, Sterns RH, Nigwekar SU, Cappuccio JD. Mortality and serum sodium: do patients die from or with hyponatremia? Clin J Am Soc Nephrol. 2011;6(5):960-965.
http://dx.doi.org/10.1016/j.mayocp.2015.07.014
1355
Sodium Correction Practice and Clinical Outcomes in Profound Hyponatremia Pierce Geoghegan, MB BCh; Andrew M. Harrison, BS; Charat Thongprayoon, MD; Rahul Kashyap, MD; Adil Ahmed, MBBS; Yue Dong, MD; Alejandro A. Rabinstein, MD; Kianoush B. Kashani, MD; and Ognjen Gajic, MD, MSc Abstract Objectives: To assess the epidemiology of nonoptimal hyponatremia correction and to identify associated morbidity and in-hospital mortality. Patients and Methods: An electronic medical record search identified all patients admitted with profound hyponatremia (sodium <120 mmol/L) from January 1, 2008, through December 31, 2012. Patients were classified as having optimally or nonoptimally corrected hyponatremia at 24 hours after admission. Optimal correction was defined as sodium correction in 24 hours of 6 through 10 mmol/L. We investigated the association between sodium correction and demographic and outcome variables, including occurrence of osmotic demyelination syndrome (ODS). Baseline characteristics by correction outcome categories were compared using the Kruskal-Wallis test for continuous variables and the c2 test for categorical variables. Odds ratios for in-hospital mortality between groups were assessed using logistic regression. Adjusted differences in hospital length of stay (LOS) and intensive care unit (ICU) LOS were assessed using the Dunnett 2-tailed t test. Results: A total of 412 patients satisfied inclusion criteria of whom 174 (42.2%) were admitted to the ICU. A total of 211 (51.2%) had optimal correction of their hyponatremia at 24 hours, 87 (21.1%) had undercorrected hyponatremia, and 114 (27.9%) had overcorrected hyponatremia. Both patient factors and treatment factors were associated with nonoptimal correction. There was a single case of ODS. Overcorrection was not associated with in-hospital mortality or ICU LOS. When adjusted for patient factors, undercorrection of profound hyponatremia was associated with an increase in hospital LOS (9.3 days; 95% CI, 1.9-16.7 days). Conclusion: Nonoptimal correction of profound hyponatremia is common. Fortunately, nonoptimal correction is associated with serious morbidity only infrequently. ª 2015 Mayo Foundation for Medical Education and Research
For editorial comment, see page 1320 From the Multidisciplinary Epidemiology and Translational Research in Intensive Care (P.G., A.M.H., C.T., R.K., A.A., Y.D., A.A.R., K.B.K., O.G.), Medical Scientist Training Program (A.M.H.), Department of Anesthesiology (R.K.), Division of Pulmonary and Critical Care Affiliations continued at the end of this article.
1348
H
yponatremia is a common electrolyte abnormality in hospitalized patients,1 and dysnatremia is an independent risk factor for hospital mortality.2 Profound hyponatremia (serum sodium <120 mmol/L) can lead to cerebral edema, and correction of the hyponatremia is necessary to prevent or treat associated neurologic complications or even death.3 Cerebral edema is a particular risk when the hyponatremia is acute (48 hours) because the brain has not yet had sufficient time to make adaptations to the hypotonic environment.4,5 In chronic hyponatremia (>48 hours), the brain has usually made adaptations to the hypotonic environment, and overly rapid
n
Mayo Clin Proc. 2015;90(10):1348-1355
correction of profound hyponatremia is associated with osmotic demyelination syndrome (ODS), which may itself cause substantial morbidity or even mortality.6 Thus, the duration of hyponatremia, if known accurately, can determine whether the primary risk to the patient is from cerebral edema or ODS. Experts have disagreed on exactly what constitutes optimal correction.7 Deciding what constitutes optimal correction is a particular challenge in patients who have hyponatremia at admission (rather than develop it during hospitalization) because the duration of hyponatremia is usually not known.8 Therefore, it can be argued that sodium correction in these patients
Mayo Clin Proc. n October 2015;90(10):1348-1355 n http://dx.doi.org/10.1016/j.mayocp.2015.07.014 www.mayoclinicproceedings.org n ª 2015 Mayo Foundation for Medical Education and Research
SODIUM CORRECTION AND OUTCOMES IN HYPONATREMIA
must account for the risk of cerebral edema (if acute) or the risk of ODS (if chronic). There is good evidence to suggest that raising the serum sodium level by a modest 4 to 6 mmol/L is more than sufficient to treat clinically significant neurologic symptoms due to profound hyponatremia.9 Data from case reports and case series suggest that there is probably a continuum of risk of ODS in chronic hyponatremia that increases when correction is faster than 8 mmol/L per 24 hours and is highest if correction rates exceed 12 to 14 mmol/L per 24 hours.10-12 The most recent guidelines recommend a ceiling of correction of 10 mmol per 24 hours or less when there is a risk of ODS and an initial correction target of 5 mmol per 24 hours to treat any symptoms of cerebral edema.8 Treatment response is unpredictable in profound hyponatremia.13,14 Nevertheless, the epidemiology of nonoptimal correction of profound hyponatremia is not well described. In particular, the risk factors for nonoptimal correction and its implications for patient outcomes have been poorly delineated. We hypothesized that nonoptimal correction of profound hyponatremia would be common and associated with increased morbidity and in-hospital mortality. In the current study, our objective was to assess the epidemiology of nonoptimal correction in patients at Mayo Clinic’s campus in Rochester, Minnesota, and to attempt to identify any associated morbidity (including ODS) and mortality. PATIENTS AND METHODS Study Design and Setting Our study was a retrospective analysis of our longstanding electronic medical record (EMR) system at the Rochester campus of Mayo Clinic. The local institutional review board reviewed and approved the project. We performed a search of the EMR to identify all patients admitted to our institution from January 1, 2008, through December 31, 2012, with a serum sodium level less than 120 mmol/L after correction for serum glucose using a well-described correction factor.15 We excluded patients younger than 18 years, patients who did not provide research authorization, those who were diagnosed as having pseudohyponatremia, and, finally, Mayo Clin Proc. n October 2015;90(10):1348-1355 www.mayoclinicproceedings.org
n
those who died in the emergency department or discharged against medical advice in the first 24 hours of treatment. When patients were admitted with profound hyponatremia more than once during the study period, only the first admission was considered in the final analysis. Study Participants and Data Collection For each patient included in the study, we further identified all serum sodium measurements undertaken in the first 72 hours of admission. Taking the time of the initial sodium check (Na0) identifying profound hyponatremia as time zero (T0), we estimated serum sodium levels at the 24-hour mark (Na24). To do this, we used the formula: Na24 ¼ Naa þ [(Nab Naa) (24 Ta)/Tb Ta)], where Naa indicates the closest serum sodium measurement before the 24-hour mark; Nab, the closest serum sodium measurement after the 24-hour mark; Ta, the time at which measurement Naa was taken; and Tb, the time at which measurement Nab was taken. The degree of correction at 24 hours (DNa24) was calculated as Na24 e Na0. Patients were judged to be optimally corrected if their DNa24 levels were 6 through 10 mmol/L. All other patients were classified as having nonoptimally corrected hyponatremia. Patients whose hyponatremia was nonoptimally corrected were further divided into those with undercorrected hyponatremia (DNa24 5 mmol/L) and overcorrected hyponatremia (DNa24 >10 mmol/L). To establish whether individual patients had experienced clinically or radiologically confirmed ODS, we performed an electronic search of the EMR to identify terms in the treating or consulting physicians notes that suggested suspicion or confirmation of a diagnosis of ODS during the admission with profound hyponatremia or up to 6 months thereafter. These terms included central pontine myelinolysis, extrapontine myelinolysis, and osmotic demyelination syndrome, as well as combinations of these. Individual medical records that contained these terms were then manually reviewed by a clinician to determine whether a diagnosis of ODS had indeed been confirmed clinically or radiologically. The records of a random sample of 10% of all patients whose EMRs did not contain any of these terms were also manually
http://dx.doi.org/10.1016/j.mayocp.2015.07.014
1349
MAYO CLINIC PROCEEDINGS
All patients admitted between January 1, 2008, through December 31, 2012, were considered for eligibility
540 Admission serum sodium <120 mmol/L
128 Excluded : • 48 Age <18 years at admission • 50 Absence of research authorization • 26 Pseudohyponatremia • 4 Discharged against medical advice or died within 24 hours 412 Met all inclusion criteria
Classified according to estimated serum sodium at 24 hours
87 Undercorrection (≤5 mmol/L at 24 hours)
211 Optimal correction (6-10 mmol/L at 24 hours)
114 Overcorrection (>10 mmol/L at 24 hours)
FIGURE. Patient selection process and subsequent categorization according to degree of correction of profound hyponatremia 24 hours after the initial sodium measurement.
searched to ensure that cases of ODS were not being missed. We additionally cross-referenced our search with a previously published cohort of ODS cases from our institution to ensure that no cases of ODS in our cohort had been missed.16 A similar strategy was used to identify instances where clinicians had identified one of a (nonexhaustive) list of common causes of hyponatremia (eg, hypovolemia, thiazide diuretic use) on initial assessment of each patient and to identify the administration of therapies, such as hypertonic saline or desmopressin. The EMR was also used to identify other relevant data, including demographic data and clinical outcome measures, using previously validated methods and the Charlson definitions for individual comorbidities, including congestive heart failure, chronic kidney disease, and cirrhosis.17,18 Primary outcomes included the epidemiology of nonoptimal correction, the incidence of clinically or radiologically confirmed ODS, in-hospital mortality, and length of stay (LOS). Secondary outcomes were the association 1350
Mayo Clin Proc.
n
between patient and treatment factors and risk of nonoptimal correction. Statistical Analyses Patient demographic characteristics were summarized using median and interquartile range (IQR) for continuous variables and number and percentage for categorical variables. The baseline characteristics by sodium outcome categories were compared using the Kruskal-Wallis test for continuous variables and the c2 test for categorical variables. The adjusted and unadjusted odds ratios for in-hospital mortality in patients with undercorrected, optimally corrected, and overcorrected hyponatremia were assessed using a logistic regression with the in-hospital mortality for the patients with optimally corrected hyponatremia being used as the referent group. The independent variables used in this multivariable logistic regression were hypertonic saline therapy, age, Charlson comorbidity score, congestive heart failure, cirrhosis or chronic liver disease, moderate or severe chronic kidney disease, and baseline sodium level. The unadjusted
October 2015;90(10):1348-1355
n
http://dx.doi.org/10.1016/j.mayocp.2015.07.014 www.mayoclinicproceedings.org
Characteristic Age, y No. of participants Female Male Total BMI Charlson comorbidity score Admission sodium, mmol/L Admission serum osmolarity, mOsmol/kg
n
http://dx.doi.org/10.1016/j.mayocp.2015.07.014
Admission urine osmolarity, mOsmol/kg Admission urine sodium, mmol/L Patients with admission urine sodium <20 mmol/L Patients with admission potassium <3 mEq/L Seizures Congestive heart failure Diagnosis of cirrhosis or chronic liver disease Moderate to severe chronic kidney disease Acute kidney injury within 24 hours Diagnosis of alcoholism at admission Taking a thiazide diuretic before admission Taking a SSRI before admission Diagnosis of hypovolemia at admission Diagnosis of pneumonia at admission SIADH in admission differential diagnosis Admitted to ICU 3-Hour APACHE III score Total fluid in, L Total fluid out, L Hypertonic saline therapy Desmopressin therapy Vasopressin receptor antagonist therapy Number of sodium checks DNR/DNI order in place Nephrology consultation Received dialysis
Undercorrection (n¼87)
Optimal correction (n¼211)
Overcorrection (n¼114)
Total (N¼412)
P value
69 (55-79) (n¼86)
71 (60-81) (n¼210)
63 (51-78) (n¼114)
69 (56-80) (n¼410)
.009
52 (59.8) 35 (40.2) 87 28 (21-33) (n¼87) 6 (4-8) (n¼87) 117 (115-119) (n¼87) 248 (243-254) (n¼58) 58 363 (291-466) (n¼71) 39 (23-66) (n¼26) 5 (19.2) (n¼26) 3 (3.5) (n¼86) 1 (1.1) (n¼87) 41 (47.1) (n¼87) 16 (18.4) (n¼87) 14 (16.1) (n¼87) 11 (12.6) (n¼87) 12 (13.8) (n¼87) 16 (18.4) (n¼87) 19 (21.8) (n¼87) 25 (28.7) (n¼87) 19 (21.8) (n¼87) 20 (23.0) (n¼87) 27 (31.0) (n¼87) 56 (37-63) (n¼27) 1.1 (0.7-2.5) (n¼39) 1.3 (0.7-2.2) (n¼39) 4 (4.6) (n¼87) 0 (0.0) (n¼87) 1 (1.1) (n¼87) 3 (2-5) (n¼87) 15 (17.2) (n¼87) 17 (19.5) (n¼87) 0 (0) (n¼87)
123 (58.3) 88 (41.7) 211 25 (22-30) (n¼206) 5 (3-8) (n¼211) 117 (114-119) (n¼211) 250 (241-258) (n¼138) 138 338 (266-429) (n¼180) 47 (21-80) (n¼67) 13 (19.4) (n¼67) 24 (11.4) (n¼211) 10 (4.7) (n¼211) 76 (36.0) (n¼211) 10 (4.7) (n¼211) 31 (14.7) (n¼211) 44 (20.9) (n¼211) 38 (18.0) (n¼211) 46 (21.8) (n¼211) 54 (25.6) (n¼211) 79 (37.4) (n¼211) 56 (26.5) (n¼211) 55 (26.0) (n¼211) 91 (43.1) (n¼211) 52 (42-68) (n¼91) 3.0 (1.4-4.2) (n¼105) 2.0 (1.2-3.0) (n¼105) 11 (5.2) (n¼211) 2 (0.9) (n¼211) 1 (0.5) (n¼211) 5 (3-7) (n¼211) 42 (19.9) (n¼211) 40 (19.0) (n¼211) 5 (2.3) (n¼211)
65 (57.0) 49 (43.0) 114 24 (22-29) (n¼112) 5 (3-8) (n¼114) 115 (112-117) (n¼114) 249 (239-258) (n¼74) 74 304 (199-384) (n¼91) 39 (19-65) (n¼29) 8 (27.6) (n¼29) 21 (18.4) (n¼114) 7 (6.1) (n¼114) 20 (17.5) (n¼114) 3 (2.6) (n¼114) 7 (6.1) (n¼114) 20 (17.5) (n¼114) 17 (14.9) (n¼114) 29 (25.4) (n¼114) 26 (22.8) (n¼114) 40 (35.1) (n¼114) 25 (21.9) (n¼114) 29 (25.4) (n¼114) 56 (49.1) (n¼114) 55 (38-72) (n¼56) 4.4 (2.4-8.3) (n¼72) 3.9 (2.3-6.5) (n¼72) 14 (12.3) (n¼114) 2 (1.8) (n¼114) 1 (0.9) (n¼114) 6 (4-8) (n¼114) 10 (8.8) (n¼114) 13 (11.4) (n¼114) 2 (1.7) (n¼114)
240 (58.3) 172 (41.7) 412 25 (22-30) (n¼405) 5 (3-7) (n¼412) 116 (114-118) (n¼412) 249 (242-257) (n¼270) 270 332 (256-427) (n¼342) 43 (21-73) (n¼122) 26 (21.3) (n¼122) 48 (11.7) (n¼411) 18 (4.4) (n¼412) 137 (33.3) (n¼412) 29 (7.0) (n¼412) 52 (12.6) (n¼412) 75 (18.2) (n¼412) 67 (16.3) (n¼412) 91 (22.1) (n¼412) 99 (24.0) (n¼412) 144 (34.9) (n¼412) 100 (24.2) (n¼412) 104 (25.2) (n¼412) 174 (42.2) (n¼412) 53 (41-68) (n¼174) 2.8 (1.3-5.1) (n¼216) 2.2 (1.3-3.9) (n¼216) 29 (7.0) (n¼412) 4 (1.0) (n¼412) 3 (0.7) (n¼412) 5 (3-7) (n¼412) 67 (16.2) (n¼412) 70 (17.0) (n¼412) 7 (1.7) (n¼412)
.93
.13 .002 <.001 .66 .003 .76 .64 .005 .21 <.001 <.001 .04 .24 .60 .49 .74 .36 .55 .86 .03 .82 <.001 <.001 .04 .45 .80 <.001 .03 .17 .35
APACHE ¼ Acute Physiology and Chronic Health Evaluation; BMI ¼ body mass index; DNR/DNI ¼ do not resuscitate/do not intubate; ICU ¼ intensive care unit; SIADH ¼ syndrome of inappropriate antidiuretic hormone secretion; SSRI ¼ selective serotonin reuptake inhibitor. b Data are presented as median (interquartile range) for continuous variables and number (percentage) for categorical variables. a
SODIUM CORRECTION AND OUTCOMES IN HYPONATREMIA
Mayo Clin Proc. n October 2015;90(10):1348-1355 www.mayoclinicproceedings.org
TABLE 1. Association of Patient and Treatment Factors to Sodium Correction Outcome at 24 Hoursa,b
1351
MAYO CLINIC PROCEEDINGS
TABLE 2. Primary and Secondary Outcome Measuresa,b Outcome measure
Undercorrection (n¼87)
Osmotic demyelination syndrome 0 (0) Survival to discharge 78 (88) Odds ratio (95% CI) of death at discharge Unadjusted 2.1 (0.8-5.3) Adjustedc 2.2 (0.8-5.6) In-hospital length of stay, d 5 (3-8) ICU length of stay, d 2 (1-4) Between group differences in hospital length of stay, d Unadjusted 9 (2-17) Adjustedc 9 (2-17) Between-group difference in ICU length of stay, d Unadjusted 1 (1 to 2) Adjusted 1 (1 to 2)
Optimal correction (n¼211) 0 (0) 200 (95) 1.0 [Reference] 1.0 [Reference] 5 (3-8) 2 (1-3)
Overcorrection (n¼114)
Total P (N¼412) value
1 (1) 108 (95)
1 (0) 386 (94)
1.0 1.0 4 2
(0.4-2.8) (0.4-3.1) (2-8) (1-3)
. . 5 (3-8) 2 (1-3)
.27 .22 .23 .24 .10 .98
0 [Reference] 0 [Reference]
1 (6 to 7) 1 (6 to 7)
. .
.02 .04
0 [Reference] 0 [Reference]
1 (1 to 2) 1 (1 to 2)
. .
.4 .58
ICU ¼ intensive care unit. Data are reported as median (interquartile range) or median (range) for continuous variables and number (percentage) for categorical variables. Between group differences in hospital and ICU length of stay are reported as difference in days (95% confidence interval for difference). c Adjusted for use of hypertonic saline, age at admission, Charlson comorbidity score, congestive cardiac failure, cirrhosis or chronic liver disease, moderate to severe chronic kidney disease, and corrected admission serum sodium level. a
b
and adjusted difference in the hospital LOS and intensive care unit (ICU) LOS was assessed using Dunnett 2-tailed t test using the LOS for patients with optimally corrected hyponatremia as the referent group. An adjusted analysis was used to reduce confounding due to being a retrospective study. The Wilson score method was used to calculate a 95% CI for the incidence of ODS in the overcorrection group. P < .05 was considered statistically significant. Statistical analyses were performed using JMP statistical software (SAS Institute Inc). RESULTS The Figure shows how patients were identified for inclusion in the study, as well as their subsequent grouping according to degree of sodium correction at 24 hours. A total of 412 patients were included; 211 patients (51.2%) had optimally corrected hyponatremia at 24 hours, 87 patients (21.1%) had undercorrected hyponatremia, and 114 (27.7%) had overcorrected hyponatremia according to our predefined criteria. The closest sodium measurements to the 24-hour mark (used to estimate Na24) were a median of 3.6 hours before (IQR hyponatremia, 1.6-7.5) and 3.7 hours after (IQR, 1.5-9.0) the 24-hour mark. Table 1 lists the patient characteristics and treatment factors and their association with the adequacy of sodium correction outcome at 24 1352
Mayo Clin Proc.
n
hours. Overall, the median age in the study population was 69 years, and 240 patients (58.2%) were female. Median serum sodium on admission after correcting for blood glucose level was 116 mmol/L. In total, 174 patients (42.2%) were admitted to an ICU. Patients were more likely to experience overcorrection if they were younger, had lower admission serum sodium values, experienced seizures, required ICU admission, or received hypertonic saline (Table 1). Of importance, patients who were admitted to the ICU were more likely to have experienced seizures than those treated on a ward (9.2% vs 0.8%, P<.001) and were also more likely to receive hypertonic saline (10.3% vs 4.6%, P¼.02). This could mean that more symptomatic patients were admitted to the ICU and required more aggressive correction. Overcorrection was also associated with larger fluid input and with larger fluid output in the first 24 hours of treatment and with lower urine osmolality at first measurement. Patients were more likely to experience undercorrection if they had higher Charlson comorbidity scores and specific comorbidities, including congestive heart failure, cirrhosis or chronic liver disease, or moderate to severe chronic kidney disease. Table 2 gives a summary of the major primary outcomes, again grouped according to degree of 24-hour sodium correction. A single
October 2015;90(10):1348-1355
n
http://dx.doi.org/10.1016/j.mayocp.2015.07.014 www.mayoclinicproceedings.org
SODIUM CORRECTION AND OUTCOMES IN HYPONATREMIA
case of ODS was observed in the overcorrected group, corresponding to an incidence of approximately 1% (95% CI, 0%-5%) in those who were overcorrected in the first 24 hours of treatment. This patient was a 40-year-old woman with a diagnosis of psychogenic polydipsia who presented with a serum sodium level of 97 mmol/L and experienced a sodium correction of 21 mmol/L in 24 hours. The patient did not have a diagnosis of alcoholism and did not experience any hypoxia in the first 24 hours of admission. In an adjusted analysis, undercorrection at 24 hours was associated with a statistically significant increase in hospital LOS of 9 days (95% CI, 2-17 days). Nonoptimal correction was not associated with a statistically significant change in ICU LOS or in hospital mortality. DISCUSSION Our results indicate that nonoptimal correction of hyponatremia was very common in patients admitted with profound hyponatremia, occurring in nearly half of the patients in this cohort. Although it was frequent, overcorrection was associated with only a single case of ODS (corresponding to an incidence of approximately 1% in those who were overcorrected) and was not associated with changes in hospital mortality or LOS. Hospital LOS was increased in patients who met our definition of undercorrection 24 hours after admission. The high incidence of overcorrection is perhaps not surprising when one considers that response to treatment in profound hyponatremia is known to be unpredictable.14 In addition, the incidence of overcorrection in our cohort is similar to that which was observed in a similar cohort, when adjusting for a different definition of overcorrection.19 This finding suggests that this is a widespread problem rather than one limited to our hospital. That said, we admit that other institutions have described lower rates of overcorrection, albeit in selected cohorts.20 Overcorrection was more common in patients admitted to the ICU. This may be because patients with symptoms of cerebral edema (eg, seizures) were more likely to be admitted to the ICU and subsequently receive aggressive correction therapy, including the use of hypertonic saline. Overcorrection was less common in patients who had higher Charlson comorbidity scores or who had active do-not-resuscitate or Mayo Clin Proc. n October 2015;90(10):1348-1355 www.mayoclinicproceedings.org
n
do-not-intubate orders. It is possible to speculate that this was because these patients were less likely to have aggressive treatment in general, including less aggressive correction of hyponatremia. Our observed incidence of ODS in those who had overcorrected hyponatremia was approximately 1% (95% CI, 0%-5%), much lower than the rate of 11% reported in the aforementioned study, which had a similar rate of overcorrection.19 The median sodium correction in the overcorrected group was 14 mmol/L (IQR, 12-16 mmol/L) so that it is possible that this is partly related to the fact that often the targets were only marginally missed in our cohort. Indeed, the only patient in our cohort who developed ODS had a rate of correction that was 2-fold faster than the maximum recommended. Clinicians should also bear in mind that the risk of ODS increases with faster rates of correction and greater degree of correction and is further modified by patient factors such as chronicity of hyponatremia, underlying nutritional status, history of alcoholism, liver transplantation, and presence of simultaneous hypoxia or sepsis.16,21,22 Some of these risk factors (eg, hypoxia) are modifiable, and some of the risk of ODS in our cohort may have been mitigated by good multidisciplinary care. Clinicians should remember that even if the risk of ODS is low, it can result in severe morbidity when it occurs. Aggressive correction is also ill-advised because an increase in the serum sodium of 4 to 6 mmol/L per 24 hours generally suffices to reverse any neurologic symptoms, and no benefit has been found with faster correction.23 Undercorrection was associated with a clinically and statistically significant increase in overall hospital LOS of 9.3 days (95% CI, 1.9-16.7 days). This was true even after adjusting for confounders that are traditionally thought to explain the association between increased morbidity and in-hospital mortality and hyponatremia, including age, diagnosis of congestive heart failure, liver disease, chronic kidney disease, or a diagnosis of malignant disease.24 We believe that the most plausible explanation for this association is that treating clinicians kept patients in the hospital longer to correct the hyponatremia, as opposed to the hyponatremia itself producing genuine morbidity. Nevertheless, it
http://dx.doi.org/10.1016/j.mayocp.2015.07.014
1353
MAYO CLINIC PROCEEDINGS
is possible that unrecognized or suboptimally treated cerebral edema produced genuine morbidity or that morbidity was produced through an alternate mechanism. Indeed, there is some evidence that early correction of profound hyponatremia (while maintaining correction rates lower than those associated with ODS) is associated with improved short-term outcomes, although this point has not been rigorously investigated.5 In our cohort, undercorrection at 24 hours was not associated with a statistically significant increase in the risk of in-hospital mortality, although it should be borne in mind that our study may not have been large enough to detect small differences in this end point. Our study has several limitations. First, the study design was retrospective. Second, it is a single-center study, and the findings would require confirmation in future multicenter studies. Third, our definitions of undercorrection and overcorrection are according to our interpretation of available evidence but are open to debate because experts have disagreed on what constitutes optimal correction. Fourth, because this study was observational, we had to estimate the serum sodium level at 24 hours because most patients did not have a serum sodium measurement at exactly this point in time. Fifth, sodium correction between 24 and 48 hours after admission was not analyzed and may influence the risk of ODS. Sixth, these definitions of nonoptimal correction are context specific (eg, the concept of overcorrection applies only where duration of hyponatremia is >48 hours). Similarly, the concept of undercorrection has been traditionally applied only to patients who are symptomatic (although this may not be a very practical distinction because there is increasing doubt that any of these patients are truly asymptomatic and patients with acute hyponatremia can present without clear symptoms but later develop severe symptoms within hours4,8). We considered our cohort as susceptible to the complications of both acute and chronic hyponatremia, although in reality each patient can only have been one or the other. However, we would argue that this approach is supported by recent guidelines that highlight that the duration of hyponatremia in these patients will generally be unknown.8 In addition, we did not identify 1354
Mayo Clin Proc.
n
which patients experienced symptoms that were ascribed to hyponatremia. CONCLUSION In the current study, we identified a high rate of nonoptimal sodium correction during the first 24 hours of treatment in a large unselected cohort of patients admitted to our institution with profound hyponatremia during a 5-five year period. Fortunately, this high rate of nonoptimal correction was not associated with increased in-hospital mortality and seemed to be associated with morbidity only infrequentlydODS occurred in only a single patient, corresponding to an incidence of approximately 1% in patients who had overcorrected hyponatremia. ACKNOWLEDGMENTS The authors gratefully acknowledge the entire Multidisciplinary Epidemiology and Translational Research in Intensive Care research group for their support, which made this project possible. The authors also thank Mayo Clinic’s Center for Translational Science Activities for their support. The contents of this publication are solely the responsibility of the authors and do not necessarily represent the official views of the National Institutes of Health. Affiliations (Continued from the first page of this article.): Medicine (Y.D., O.G.), Department of Neurology (A.A.R.), and Division of Nephrology and Hypertension (K.B.K.), Mayo Clinic, Rochester, MN; and Wichita Falls Family Practice Residency Program, North Central Texas Medical Foundation, Wichita Falls (A.A.). Grant Support: The work was supported by the Mayo Clinic’s Center for Translational Science Activities grant UL1 TR000135 from the National Center for Advancing Translational Sciences. Correspondence: Address to Pierce Geoghegan, MB, BCh, BAO, BA, Department of Anaesthesia and Critical Care, Rotunda Hospital, Dublin, Ireland ([email protected]).
REFERENCES 1. DeVita MV, Gardenswartz MH, Konecky A, Zabetakis PM. Incidence and etiology of hyponatremia in an intensive care unit. Clin Nephrol. 1990;34(4):163-166. 2. Darmon M, Diconne E, Souweine B, et al. Prognostic consequences of borderline dysnatremia: pay attention to minimal serum sodium change. Crit Care. 2013;17(1):R12. 3. Schrier RW, Bansal S. Diagnosis and management of hyponatremia in acute illness. Curr Opin Crit Care. 2008;14(6):627-634. 4. Arieff AI. Hyponatremia, convulsions, respiratory arrest, and permanent brain damage after elective surgery in healthy women. N Engl J Med. 1986;314(24):1529-1535.
October 2015;90(10):1348-1355
n
http://dx.doi.org/10.1016/j.mayocp.2015.07.014 www.mayoclinicproceedings.org
SODIUM CORRECTION AND OUTCOMES IN HYPONATREMIA
5. Nzerue CM, Baffoe-Bonnie H, You W, Falana B, Dai S. Predictors of outcome in hospitalized patients with severe hyponatremia. J Natl Med Assoc. 2003;95(5):335-343. 6. Karp BI, Laureno R. Pontine and extrapontine myelinolysis: a neurologic disorder following rapid correction of hyponatremia. Medicine (Baltimore). 1993;72(6):359-373. 7. Overgaard-Steensen C. Initial approach to the hyponatremic patient. Acta Anaesthesiol Scand. 2011;55(2):139-148. 8. Spasovski G, Vanholder R, Allolio B, et al. Clinical practice guideline on diagnosis and treatment of hyponatraemia. Nephrol Dial Transplant. 2014;29(suppl 2):i1-i39. 9. Sterns RH, Hix JK, Silver S. Treatment of hyponatremia. Curr Opin Nephrol Hypertens. 2010;19(5):493-498. 10. Adrogué HJ, Madias NE. Hyponatremia. N Engl J Med. 2000; 342(21):1581-1589. 11. Sterns RH, Nigwekar SU, Hix JK. The treatment of hyponatremia. Semin Nephrol. 2009;29(3):282-299. 12. Brunner JE, Redmond JM, Haggar AM, Kruger DF, Elias SB. Central pontine myelinolysis and pontine lesions after rapid correction of hyponatremia: a prospective magnetic resonance imaging study. Ann Neurol. 1990;27(1):61-66. 13. Vaidya C, Ho W, Freda BJ. Management of hyponatremia: providing treatment and avoiding harm. Cleve Clin J Med. 2010;77(10):715-726. 14. Nguyen MK. Quantitative approaches to the analysis and treatment of the dysnatremias. Semin Nephrol. 2009;29(3):216-226. 15. Hillier TA, Abbott RD, Barrett EJ. Hyponatremia: evaluating the correction factor for hyperglycemia. Am J Med. 1999;106(4): 399-403.
Mayo Clin Proc. n October 2015;90(10):1348-1355 www.mayoclinicproceedings.org
n
16. Graff-Radford J, Fugate JE, Kaufmann TJ, Mandrekar JN, Rabinstein AA. Clinical and radiologic correlations of central pontine myelinolysis syndrome. Mayo Clin Proc. 2011;86(11):1063-1067. 17. Ahmed A, Thongprayoon C, Pickering BW, et al. Towards prevention of acute syndromes: electronic identification of at-risk patients during hospital admission. Appl Clin Inform. 2014;5(1):58-72. 18. Singh B, Singh A, Ahmed A, et al. Derivation and validation of automated electronic search strategies to extract Charlson comorbidities from electronic medical records. Mayo Clin Proc. 2012;87(9):817-824. 19. Vu T, Wong R, Hamblin PS, Zajac J, Grossmann M. Patients presenting with severe hypotonic hyponatremia: etiological factors, assessment, and outcomes. Hosp Pract (1995). 2009;37(1):128-136. 20. Sood L, Sterns RH, Hix JK, Silver SM, Chen L. Hypertonic saline and desmopressin: a simple strategy for safe correction of severe hyponatremia. Am J Kidney Dis. 2013;61(4):571-578. 21. Martin RJ. Central pontine and extrapontine myelinolysis: the osmotic demyelination syndromes. J Neurol Neurosurg Psychiatry. 2004;75(suppl 3):iii22-iii28. 22. Singh TD, Fugate JE, Rabinstein AA. Central pontine and extrapontine myelinolysis: a systematic review. Eur J Neurol. 2014;21(12):1443-1450. 23. Verbalis JG, Goldsmith SR, Greenberg A, Schrier RW, Sterns RH. Hyponatremia treatment guidelines 2007: expert panel recommendations. Am J Med. 2007;120(11 suppl 1):S1-S21. 24. Chawla A, Sterns RH, Nigwekar SU, Cappuccio JD. Mortality and serum sodium: do patients die from or with hyponatremia? Clin J Am Soc Nephrol. 2011;6(5):960-965.
http://dx.doi.org/10.1016/j.mayocp.2015.07.014
1355