- Email: [email protected]
Effect of Hypernatremia on Outcomes After severe Traumatic Brain Injury: A Nationwide Inpatient Sample analysis
Original Article
Effect of Hypernatremia on Outcomes After severe Traumatic Brain Injury: A Nationwide Inpatient Sample analysis Haydn Hoffman, Muhammad S. Jalal, Lawrence S. Chin
OBJECTIVE: Induced hypernatremia is frequently used to reduce intracranial pressure in patients with severe traumatic brain injury (TBI). This technique is controversial, and some studies have independently associated hypernatremia with worse outcomes after TBI. We sought to investigate this potential association in a large healthcare database.
-
METHODS: The Nationwide Inpatient Sample was used to obtain data on all adults who had been discharged from 2002 to 2011 with a primary diagnosis of TBI who required mechanical ventilation, intracranial pressure monitoring, or craniotomy/craniectomy. Patients with diabetes insipidus were excluded. The patients with hypernatremia were assigned to the hypernatremia group, and the rest were assigned to the control group. The primary outcome was inhospital mortality, and the secondary outcomes included the length of stay, nonroutine hospital discharge, total hospital charges, tracheostomy, and gastrostomy placement.
hypernatremia also had greater rates of tracheostomy and gastrostomy placement. CONCLUSIONS: Hypernatremia was associated with poorer outcomes in patients with severe TBI. This finding warrants further investigation in a prospective, randomized study.
-
-
RESULTS: A total of 85,579 patients without a diagnosis of hypernatremia (control group) and 4542 patients with a diagnosis of hypernatremia (hypernatremia group) were identified. When controlling for age, comorbidities, gender, and cerebral edema, hypernatremia was associated with an increased rate of in-hospital mortality (odds ratio, 1.51; 95% confidence interval, 1.39e1.65), a longer mean length of stay (23.65 vs. 12.12 days; P < 0.001), an increased rate of nonroutine hospital discharge (odds ratio, 2.58; 95% confidence interval, 2.28e2.92), and greater mean total hospital cost ($227,112 vs. $112,507; P < 0.001). The patients with
-
Key words Hypernatremia - Nationwide inpatient sample - Traumatic brain injury -
Abbreviations and Acronyms DI: Diabetes insipidus EVD: External ventricular drain GCS: Glasgow Coma Scale ICP: Intracranial pressure LOS: Length of stay NIS: Nationwide Inpatient Sample
INTRODUCTION
S
evere traumatic brain injury (TBI) has the potential to disrupt the bloodebrain barrier and cause egress of fluid into the intracellular and extracellular spaces, resulting in cerebral edema and increased intracranial pressure (ICP).1 Induced hypernatremia is frequently used as a method of reducing acute elevations in ICP after TBI.2,3 Hypernatremia is a physiological condition in which the homeostatic water balance is disrupted, resulting in an increased concentration of serum sodium. An increase in serum sodium concentration is believed to cause a decrease in cell volume and cerebral blood volume, thereby decreasing cerebral edema. It also can modulate neuroinflammatory pathways, restore neuronal membrane potentials, and reduce blood viscosity.4,5 For these reasons, hypertonic saline is frequently given as a bolus to induce short periods of hyperosmolarity after TBI.6 However, hypernatremia can occur as a result of repeated dosing. Although many have endorsed the therapeutic effects of hyperosmolar therapy, others have shown that hypernatremia after severe TBI is independently associated with poor patient
PEG: Percutaneous endoscopic gastrostomy TBI: Traumatic brain injury Department of Neurosurgery, State University of New York Upstate Medical University, Syracuse, New York, USA To whom correspondence should be addressed: Haydn Hoffman, M.D. [E-mail: [email protected]] Citation: World Neurosurg. (2018). https://doi.org/10.1016/j.wneu.2018.07.089 Journal homepage: www.WORLDNEUROSURGERY.org Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2018 Elsevier Inc. All rights reserved.
WORLD NEUROSURGERY -: e1-e7, - 2018
www.WORLDNEUROSURGERY.org
e1
ORIGINAL ARTICLE HAYDN HOFFMAN ET AL.
HYPERNATREMIA AFTER TRAUMATIC BRAIN INJURY
prognosis and death.7 Hypernatremia has also been associated with poorer outcomes in patients with mild TBI.8 It remains controversial whether hypernatremia independently worsens the outcomes or is simply a surrogate marker for illness severity. For example, severe TBI can be associated with central diabetes insipidus (DI), a potentially confounding variable that causes hypernatremia and independently increases mortality.7 The use of iatrogenic hypernatremia after TBI remains controversial and its effects on patient prognosis continue to be debated. This is reflected in the most recent guidelines for severe TBI, which do not offer a specific recommendation regarding the use of hyperosmolar therapy.9 Thus, we sought to assess the effect of hypernatremia on the outcomes of patients with severe TBI using a large national healthcare database. METHODS The nationwide inpatient sample (NIS) was used to identify the patients for present study. The NIS is the largest publicly available all-payer healthcare database. It includes inpatient data obtained from hospital discharges in all states participating in the Healthcare Cost and Utilization Project. Approximately 1000 hospitals are sampled annually to include 8 million discharges for a 20% stratified sample of U.S. community hospitals. Discharge weights are provided to obtain national estimates of disease incidence. The data include 15 diagnoses and procedures per discharge, provided as International Classification of Diseases, Ninth Revision, Clinical Modification codes. Patients aged 18 years who had been discharged from 2002 to 2011 with a primary discharge diagnosis of TBI were identified. The Clinical Classifications Software code 233 was used to identify patients with TBI. The Clinical Classifications Software code 233 encompasses several International Classification of Diseases, Ninth Revision, Clinical Modification codes for TBI (800.00e801.9 and 803.1e804.9 codes, fracture of cranial vault, skull base, or facial bone with intracranial injury; 850.0e854.1 codes, concussion, cerebral contusion, subdural hematoma, epidural hematoma, other and unspecified traumatic intracranial hemorrhage). Patients with a primary diagnosis of V1552 (history of TBI), elective hospital admission, or a secondary diagnosis of DI (code 253.5) were excluded. Because the presenting Glasgow Coma Scale (GCS) score was not available, patients with severe TBI were identified as those undergoing external ventricular drain placement (EVD; code 02.2), ICP monitor placement (code 01.10), craniectomy or craniotomy (code 01.24), hematoma drainage (codes 01.31 or 01.39), or mechanical ventilation (codes 96.04, 96.70, 96.71, 96.72). Those who had not undergone any of these procedures were excluded. Patients with a diagnosis of hypernatremia (code 276.0) were included in the hypernatremia group and those without this diagnosis were included in the control group of the present study. We did not have access to data regarding the duration, degree, or etiology of hypernatremia. The primary outcome in the present study was in-hospital mortality, and the secondary outcomes included the length of stay (LOS), nonroutine hospital discharge, total hospital charges, tracheostomy (codes 31.1, 31.21, 31.29), and percutaneous endoscopic gastrostomy (PEG; code 43.11). With the exception of the last 2 outcomes, these items are provided as discrete data
e2
www.SCIENCEDIRECT.com
elements for each hospital discharge in the NIS. The NIS classifies discharge disposition as routine, transfer to a short-term hospital, transfer to a skilled nursing facility or an intermediate care facility, and home healthcare. These were dichotomized to routine or nonroutine discharges. Age, race (classified as white, black, Hispanic, Asian or Pacific Islander, Native American, or other), sex, and 29 Elixhauser comorbidities were adjusted for in the outcome analysis. The Elixhauser comorbidity measures cover a broad range of conditions that are independently associated with increased LOS, hospital charges, and mortality and are provided in the NIS data set.10 Univariate analyses using Student’s t test for continuous variables and Fisher’s exact test for categorical variables were performed to compare the demographic data, hospital characteristics, and comorbidities in patients with and without hypernatremia. A multiple logistic regression analysis was used to compare the incidences of potential predictors of hypernatremia in the hypernatremia and control groups. Multiple logistic regression or multiple linear regression analyses were performed to compare binomial or continuous outcomes between each group, respectively. Age, sex, race, and Elixhauser comorbidities were adjusted for in each regression analysis. The presence of cerebral edema (code 348.5) was adjusted for in the outcomes analysis. Statistical analysis was performed using IBM SPSS Statistics, version 24.0 (IBM Corp., Armonk, NY). RESULTS The total number of hospital discharges from 2002 to 2011 with a primary diagnosis of TBI was 331,759. After excluding cases of nonsevere TBI and those with a secondary diagnosis of DI, 90,121 cases were included in the present study. Of these, 10,830 (12.0%) required EVD or ICP monitor placement, 38,015 (42.2%) underwent craniotomy/craniectomy with or without hematoma evacuation, and 66,606 (73.9%) required mechanical ventilation. A total of 85,579 patients (95.0%) did not have a diagnosis of hypernatremia (control group) and 4542 (5.0%) did have hypernatremia (hypernatremia group). In the univariate analyses, the hypernatremia group had a slightly older mean age and included more men and more patients treated in urban and teaching hospitals. The full list of demographic data and comorbidities is included in Table 1. Patients aged 36e65 years had a greater incidence of hypernatremia compared with patients aged <36 years. Female sex was associated with a lower odds of hypernatremia than male sex. Compared with white patients, none of the other races had greater odds of developing hypernatremia. Coagulopathy and hypernatremia were significantly associated. The complete list of the predictors of hypernatremia is presented in Table 2. The outcomes were worse in the hypernatremia group than in the control group (Table 3 and Figure 1). A binary logistic regression analysis was performed to evaluate the association between hypernatremia and in-hospital mortality. The logistic regression model was highly statistically significant (c2 ¼ 6790.35; P < 0.001). The rate of in-hospital mortality was 24.8% in the control group and 30.7% in the hypernatremia group (odds ratio, 1.51; 95% confidence interval, 1.39e1.65; P < 0.001). A multiple regression analysis was run to predict the LOS from hypernatremia and the additional covariates. These variables significantly predicted the LOS [F(34,
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2018.07.089
ORIGINAL ARTICLE HAYDN HOFFMAN ET AL.
HYPERNATREMIA AFTER TRAUMATIC BRAIN INJURY
Table 1. Comparison of Demographic, Hospital Characteristics, and Comorbidities in Patients with Traumatic Brain Injury with and without Hypernatremia Hypernatremia (n, %) Variable Patients (n) Mean age (years)
No
Yes 4542
NA
54.59 21.66
0.001 <0.001
24,348 (28.5)
1117 (24.6)
36e65
30,581 (35.7)
1782 (39.2)
>65
30,650 (35.8)
1643 (36.2)
Male
60,972 (71.3)
3381 (74.5)
Female
24,508 (28.7)
1160 (25.5)
Race
Hypertension
Yes
1213 (1.4)
137 (3.0)
27,642 (32.7)
1620 (35.9)
Obesity
1464 (1.7)
131 (2.9)
Renal failure
3358 (4.0)
355 (7.9)
Mean number of chronic conditions
3.62 2.81
4.72 3.02
P Value
<0.001
NA, not applicable.
0.207
White
46,269 (69.1)
2615 (69.6)
Black
7243 (10.8)
426 (11.3)
Hispanic
8482 (12.7)
433 (11.5)
Asian/Pacific Islander
1843 (2.8)
94 (2.5)
Native American
575 (0.9)
38 (1.0)
Other
2541 (3.8)
153 (4.1)
Northeast
15,723 (18.4)
747 (16.4)
Midwest
17,428 (20.4)
958 (21.1)
South
33,761 (39.5)
1948 (42.9)
West
18,667 (21.8)
889 (19.6)
2927 (3.5)
148 (3.3)
<0.001
Hospital region
Hospital bed size
16,228 (19.2)
770 (17.3)
Large
65,533 (77.4)
3541 (79.4) <0.001
Location Urban
82,308 (97.2)
4377 (98.2)
Rural
2380 (2.8)
82 (1.8)
Teaching
62,579 (73.9)
3450 (77.4)
Nonteaching
22,109 (26.1)
1009 (22.6)
16,438 (19.5)
758 (16.8)
Congestive heart failure
5157 (6.1)
487 (10.8)
Chronic pulmonary disease
997 (1.2)
136 (0.2)
<0.001
Hospital teaching status
<0.001
Comorbidities
Continues
WORLD NEUROSURGERY -: e1-e7, - 2018
2780901) ¼ 327.20; P < 0.001]. The mean LOS was lower in the control group (12.12 days) than in the hypernatremia group (23.65 days; P < 0.001). A binary logistic regression analysis was also performed to evaluate the association between hypernatremia and nonroutine hospital discharge. The logistic regression model was highly statistically significant (c2 ¼ 12016.98; P < 0.001). The rate of nonroutine hospital discharge was 70.3% in the control group compared with 91% in the hypernatremia group (odds ratio, 2.58; 95% confidence interval, 2.28e2.92; P < 0.001). Hypernatremia was associated with a greater mean total hospital cost ($227,112) compared with the controls group ($112,507; P < 0.001). Both tracheostomy and PEG placement were performed more frequently in the hypernatremia group (Table 3). DISCUSSION
0.005
Medium
Alcohol abuse
Diabetes with chronic complications
No
<0.001
Sex
Small
P Value
85,579
18e35
Hypernatremia (n, %) Variable
53.44 22.71
Age group (years)
Table 1. Continued
The use of hypertonic saline to achieve hypernatremia in patients with TBI has been common practice in neurointensive care units. Noniatrogenic factors can also contribute to hypernatremia, such as hypovolemia, insensible free water losses, and DI. Hypernatremia is used therapeutically to decrease cerebral edema and ICP with the goal of reducing secondary injury.11 Compared with other therapies for intracranial hypertension such as hyperventilation and barbiturate coma, hypernatremia might be associated with fewer adverse effects. This has been challenged, however, with several studies finding a correlation between hypernatremia and worse clinical outcomes in patients with TBI.7,8,12-14 This seems plausible, because hypernatremia has been shown to be associated with worse outcomes in other intensive care unit settings.15 Given the frequency with which hypernatremia is applied in TBI, this therapy warrants close scrutiny. To the best of our knowledge, the present study is the first to use the NIS to evaluate the outcomes associated with hypernatremia after TBI. The univariate analysis revealed that the patients with hypernatremia were older and tended to have more comorbidities on admission to the hospital. Patients aged 36e65 years had a greater incidence of hypernatremia compared with those aged <36 years.
www.WORLDNEUROSURGERY.org
e3
ORIGINAL ARTICLE HAYDN HOFFMAN ET AL.
HYPERNATREMIA AFTER TRAUMATIC BRAIN INJURY
Table 2. Multivariate Analysis of Potential Predictors of Hypernatremia Development Variable
P Value
OR
95% CI
<0.001
1.19
1.09e1.30
0.074
0.91
0.82e1.01
Black
0.996
1.00
0.90e1.11
Hispanic
0.079
0.91
0.82e1.01
Asian or Pacific Islander
0.126
0.85
0.68e1.05
Native American
0.356
1.17
0.84e1.64
Other
0.567
1.05
0.89e1.25
Female sex
<0.001
0.80
0.74e0.87
Alcohol abuse
<0.001
0.78
0.71e0.86
Congestive heart failure
<0.001
1.70
1.51e1.91
Coagulopathy
<0.001
2.41
2.20e2.63
Diabetes
<0.001
1.61
1.31e1.97
Drug abuse
0.051
0.86
0.73e1.00
Hypertension
0.051
1.08
1.00e1.17
Obesity
0.003
1.37
1.12e1.68
Age (years) 36e65 >65 Race
Comorbidities
Renal failure Valvular heart disease
<0.001
1.45
1.26e1.67
0.075
0.84
0.70e1.02
CI, confidence interval; OR, odds ratio.
Furthermore, female sex was associated with a lower odds of having hypernatremia than the male sex. This is consistent with other studies illustrating a decline in total water body content in older patients and women, in particular, which could predispose them to hypernatremia.16 Similarly, of the comorbidities analyzed, we found that patients with coagulopathy and renal failure had some of the greatest odds of having hypernatremia. Increased rates of renal failure have been shown to be associated with hypernatremia,13 which might result from the kidney’s inability to maintain osmotic equilibrium. Coagulopathy was found to be significantly associated with hypernatremia. It is possible that patients treated with hypertonic fluids and osmotic diuretics were intravascularly depleted, making them more prone to consumption coagulopathies. Another possibility is that patients with hypernatremia had more severe TBI, which is known to be a risk factor for the development of coagulopathy.17 However, all the patients in present study had severe TBI and were at risk of developing an associated coagulopathy. The patients in the hypernatremia group had greater mortality rates and longer LOSs than those in the control group. The rates of tracheostomy, PEG placement, and nonroutine hospital discharge were greater in the hypernatremia group, suggesting poorer functional outcomes in those who did survive. Because
e4
www.SCIENCEDIRECT.com
severe TBI is associated with the development of central DI, this could have resulted in confounders in studies assessing the relationship between hypernatremia and outcomes.8,18,19 Thus, DI was a part of our exclusion criteria. It is possible that neurointensivists are more likely to aggressively treat severe TBI with hypernatremia rather than moderate or mild cases; thus, we excluded nonsevere injuries. The NIS does not directly specify the severity of TBI. Therefore, we selected for cases of severe TBI by identifying those who required an EVD or another form of ICP monitoring, because the guidelines for management of severe TBI recommend ICP monitoring in patients with a GCS of 8.9 This GCS threshold is also used as an indication for intubation20; thus, mechanical ventilation was added to the inclusion criteria. One could argue that the patients in the hypernatremia group were more likely to have experienced cerebral edema, which could have confounded the outcomes. However, we adjusted for the presence of cerebral edema in the analysis to account for this. Although our study is the first to use a large healthcare database to investigate this topic, it is not the first to identify worse outcomes in those with hypernatremia. In a retrospective analysis of patients with severe TBI from a single institution, Vedantam et al.8 demonstrated a greater mortality rate in those with hypernatremia. Both Li et al.12 and Aiyagari et al.13 showed that an increasing severity of hypernatremia was associated with an increasing incidence of mortality. Additionally, Shehata et al.,21 in a singleinstitution study, reported greater mortality and longer LOSs for TBI patients with hypernatremia compared with those who were normonatremic. Our mortality rate of 30.7% in the hypernatremia group is comparable to the rate of 26% reported by Maggiore et al.7 in the patients with hypernatremia from their cohort. Our rate is also similar to the overall mortality rate of 30% in hypernatremic patients reported by Aiyagari et al.13 but substantially lower than the incidence of 67% reported by Li et al.12 In the latter study, most (54%) of the patients had severe hypernatremia (>160 mEq/L), which might have contributed to their greater mortality rate. A number of reasons are possible for why hypernatremia could contribute to worse outcomes after TBI. Although hypernatremia can reduce cerebral edema, the resulting homeostatic imbalance might be deleterious toward other organ systems.22 Hypernatremia has been associated with greater rates of renal failure, which can occur through renal vasoconstriction, resulting in a decreased glomerular filtration rate.23,24 Accordingly, Froelich et al25 found that in neurocritically ill patients, continuous hypertonic saline therapy was associated with greater levels of creatinine and blood urea nitrogen. Hypernatremic states have also been shown to be associated with rhabdomyolysis.26 Hypernatremia could also have negative effects on cardiac contractility.27 Additionally, although the goal of induced hypernatremia is generally to avoid secondary brain injury, some studies have suggested it might actually contribute to this. In animal studies, hypernatremia has been shown to cause myelinolysis and cellular necrosis.28 Furthermore, attempts to correct hypernatremia too quickly can cause rebound elevations in ICP, cerebral edema, and seizures.29 The use of a nationwide data set gives our results strong generalizability. Most of the reported data evaluating the effect of hypernatremia on the outcomes after TBI have come from singleinstitution analyses. However, inherent limitations exist to the use
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2018.07.089
ORIGINAL ARTICLE HAYDN HOFFMAN ET AL.
HYPERNATREMIA AFTER TRAUMATIC BRAIN INJURY
Table 3. Outcomes of Patients with and without Hypernatremia Hypernatremia (n, %) Outcome
No
Yes
P Value
OR
95% CI
Mortality
21,200 (24.8)
1394 (30.7)
<0.001
1.51
1.39e1.65
12.12 15.71
23.65 22.74
<0.001
NA
NA
Mean LOS (days) Nonroutine disposition Mean total hospital cost ($)
60,162 (70.3)
4134 (91.0)
<0.001
2.58
2.28e2.92
112,507 138,145
227,112 196,669
<0.001
NA
NA
Tracheostomy
12,589 (14.7)
1777 (39.1)
<0.001
2.23
2.04e2.42
PEG placement
9566 (11.2)
1401 (30.8)
<0.001
1.95
1.78e2.13
CI, confidence interval; LOS, length of stay; NA, not applicable; OR, odds ratio; PEG, percutaneous endoscopic gastrostomy.
of the NIS that should be considered when interpreting these results. We did not have access to each patient’s presenting GCS; thus, the presence of severe TBI was determined by surrogate factors, such as the need for ICP monitoring or mechanical
ventilation. Given that the incidence of mechanical ventilation in patients with severe TBI nears 100% in reported cohorts,7,30-32 we believe we identified the vast majority of patients with severe TBI with data in the NIS database. The sodium level for diagnosing
Figure 1. Primary and secondary outcomes in the hypernatremia group compared with the control group (**P < 0.001).
WORLD NEUROSURGERY -: e1-e7, - 2018
www.WORLDNEUROSURGERY.org
e5
ORIGINAL ARTICLE HAYDN HOFFMAN ET AL.
HYPERNATREMIA AFTER TRAUMATIC BRAIN INJURY
hypernatremia could have varied between institutions, and the duration and etiology of hypernatremia for each patient was unknown. Furthermore, the use of a validated functional outcome measurement in TBI such as the Glasgow Outcome Scale would have been preferred. Instead, we had to use surrogate outcome measures such as tracheostomy or PEG placement and discharge disposition. It is possible that patients with nonroutine dispositions at hospital discharge eventually improved enough to regain functional independence. Also, the retrospective nature prevented us determining causation. Likewise, the controversy regarding whether hypernatremia is merely a surrogate for illness severity could not be addressed in the present study. We only included patients with severe TBI to minimize the confounding effects of injury severity. Additional information about each patient’s injury, such as the location of the lesions and the amount of cerebral
REFERENCES 1. Tan SK, Kolmodin L, Sekhon MS, Qiao L, Zou J, Henderson WR, et al. The effect of continuous hypertonic saline infusion and hypernatremia on mortality in patients with severe traumatic brain injury: a retrospective cohort study. Can J Anaesth. 2016;63:664-673. 2. Peterson B, Khanna S, Fisher B, Marshall L. Prolonged hypernatremia controls elevated intracranial pressure in head-injured pediatric patients. Crit Care Med. 2000;28:1136-1143. 3. Mangat HS, Chiu YL, Gerber LM, Alimi M, Ghajar J, Härtl R. Hypertonic saline reduces cumulative and daily intracranial pressure burdens after severe traumatic brain injury. J Neurosurg. 2015;122:202-210.
edema present, would have aided in matching the groups; however, these data are not provided in the NIS database. Despite these limitations, we identified a robust association between hypernatremia and a variety of unfavorable outcomes that merits close scrutiny of this therapy. Evaluation in a prospective, randomized study is warranted to assess for causality. Such a study will need to control for variations in the amount of cerebral edema, the severity of ICP elevation, and the number of hyperosmolar therapies administered. CONCLUSIONS After severe TBI, hypernatremia has been associated with a greater incidence of in-hospital mortality, longer LOS, and nonroutine discharge disposition. A prospective, randomized study on the effects of hypernatremia after TBI is merited.
management of severe traumatic brain injury, fourth edition. Neurosurgery. 2017;80:6-15.
patients with serious head injury. BMJ. 1993;307: 547-552.
10. Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care. 1998;36:8-27.
21. Shehata M, Ragab D, Khaled M, Hegazy M, Hussein A, Khaled H. Impact of hypernatremia on patients with sever traumatic brain injury. Crit Care. 2010;14:355.
11. Khanna S, Davis D, Peterson B, Fisher B, Tung H, O’Quigley J, et al. Use of hypertonic saline in the treatment of severe refractory posttraumatic intracranial hypertension in pediatric traumatic brain injury. Crit Care Med. 2000;28:1144-1151. 12. Li M, Hu YH, Chen G. Hypernatremia severity and the risk of death after traumatic brain injury. Injury. 2013;44:1213-1218. 13. Aiyagari V, Deibert E, Diringer MN. Hypernatremia in the neurologic intensive care unit: how high is too high. J Crit Care. 2006;21:163-172.
4. Murphy N, Auzinger G, Bernel W, Wendon J. The effect of hypertonic sodium chloride on intracranial pressure in patients with acute liver failure. Hepatology. 2004;39:464-470.
14. Kolmodin L, Sekhon MS, Henderson WR, Turgeon AF, Griesdale DE. Hypernatremia in patients with severe traumatic brain injury: a systematic review. Ann Intensive Care. 2013;3:35.
5. Muizelaar JP, Wei EP, Kontos HA, Becker DP. Cerebral blood flow is regulated by changes in blood pressure and in blood viscosity alike. Stroke. 1986;17:44-48.
15. Waite MD, Fuhrman SA, Badawi O, Zuckerman IH, Franey CS. Intensive care unitacquired hypernatremia is an independent predictor of increased mortality and length of stay. J Crit Care. 2013;28:405-412.
6. Kahle KT, Walcott BP, Simard JM. Continuous hyperosmolar therapy for traumatic brain injuryassociated cerebral edema: as good as it gets, or an iatrogenic secondary insult? J Clin Neurosci. 2013;20:30-31. 7. Maggiore U, Picetti E, Antonucci E, Parenti E, Regolisti G, Mergoni M, et al. The relation between the incidence of hypernatremia and mortality in patients with severe traumatic brain injury. Crit Care. 2009;13:R110. 8. Vedantam A, Robertson CS, Gopinath SP. Morbidity and mortality associated with hypernatremia in patients with severe traumatic brain injury. Neurosurg Focus. 2017;43:E2. 9. Carney N, Totten AM, O’Reilly C, Ullman JS, Hawryluk GW, Bell MJ, et al. Guidelines for the
e6
www.SCIENCEDIRECT.com
16. Kugler JP, Hustead T. Hyponatremia and hypernatremia in the elderly. Am Fam Physician. 2000;15: 3623-3630. 17. Laroche M, Kutcher ME, Huang MC, Cohen MJ, Manley GT. Coagulopathy after traumatic brain injury. Neurosurgery. 2012;70:1334-1345. 18. Hadjizacharia P, Beale EO, Inaba K, Chan LS, Demetriades D. Acute diabetes insipidus in severe head injury: a prospective study. J Am Coll Surg. 2008;207:477-484. 19. Boughey JC, Yost MJ, Bynoe RP. Diabetes insipidus in the head-injured patient. Am Surg. 2004;70: 500-503. 20. Gentleman D, Dearden M, Midgley S, Maclean D. Guidelines for resuscitation and transfer of
22. Lindner G, Funk GC. Hypernatremia in critically ill patients. J Crit Care. 2013;28:216. 23. Wilcox CS. Regulation of renal blood flow by plasma chloride. J Clin Invest. 1983;71:726-735. 24. Gerber JG, Branch RA, Nies AS, Hollifield JW, Gerkens JF. Influence of hypertonic saline on canine renal blood flow and renin release. Am J Physiol. 1979;237:F441-F446. 25. Froelich M, Ni Q, Wess C, Ougorets I, Härtl R. Continuous hypertonic saline therapy and the occurrence of complications in neurocritically ill patients. Crit Car Med. 2009;37:1433-1441. 26. Kung AW, Pun KK, Lam KS, Yeung RT. Rhabdomyolysis associated with cranial diabetes insipidus. Postgrad Med J. 1991;67:912-913. 27. Kozeny GA, Murdock DK, Euler DE, Hano JE, Scanlon PJ, Bansal VK, et al. In vivo effects of acute changes in osmolality and sodium concentration on myocardial contractility. Am Heart J. 1985;109:290-296. 28. Ayus JC, Armstrong DL, Arieff AI, Gabbai R, ID G. Effects of hypernatraemia in the central nervous system and its therapy in rats and rabbits. J Physiol. 1996;492:243-255. 29. De Petris L, Luchetti A, Emma F. Cell volume regulation and transport mechanisms across the blood-brain barrier: implications for the management of hypernatraemic states. Eur J Pediatr. 2001;160:71-77. 30. Khajavikhan J, Aminolah V, Khani A, Jaafarpour M, Kokhazade T. Outcome and predicting factor following severe traumatic brain injury: a retrospective cross-sectional study. J Clin Diagn Res. 2016;10:PC16-PC19.
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2018.07.089
ORIGINAL ARTICLE HAYDN HOFFMAN ET AL.
31. Godbolt AK, Stenberg M, Jakobsson J, Sorjonen K, Krakau K, Stålnacke BM, et al. Subacute complications during recovery from severe traumatic brain injury: frequency and associations with outcome. BMJ Open. 2015;5. e007208. 32. Haddad S, Aldawood AS, Alferayan A, Russell NA, Tamim HM, Arabi YM. Relationship between intracranial pressure monitoring and outcomes in
HYPERNATREMIA AFTER TRAUMATIC BRAIN INJURY
severe traumatic brain injury patients. Anaesth Intensive Care. 2011;39:1043-1050.
Citation: World Neurosurg. (2018). https://doi.org/10.1016/j.wneu.2018.07.089 Journal homepage: www.WORLDNEUROSURGERY.org
Conflict of interest statement: The authors declare that the article content was composed in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2018 Elsevier Inc. All rights reserved.
Received 12 April 2018; accepted 10 July 2018
WORLD NEUROSURGERY -: e1-e7, - 2018
www.WORLDNEUROSURGERY.org
e7
Effect of Hypernatremia on Outcomes After severe Traumatic Brain Injury: A Nationwide Inpatient Sample analysis Haydn Hoffman, Muhammad S. Jalal, Lawrence S. Chin
OBJECTIVE: Induced hypernatremia is frequently used to reduce intracranial pressure in patients with severe traumatic brain injury (TBI). This technique is controversial, and some studies have independently associated hypernatremia with worse outcomes after TBI. We sought to investigate this potential association in a large healthcare database.
-
METHODS: The Nationwide Inpatient Sample was used to obtain data on all adults who had been discharged from 2002 to 2011 with a primary diagnosis of TBI who required mechanical ventilation, intracranial pressure monitoring, or craniotomy/craniectomy. Patients with diabetes insipidus were excluded. The patients with hypernatremia were assigned to the hypernatremia group, and the rest were assigned to the control group. The primary outcome was inhospital mortality, and the secondary outcomes included the length of stay, nonroutine hospital discharge, total hospital charges, tracheostomy, and gastrostomy placement.
hypernatremia also had greater rates of tracheostomy and gastrostomy placement. CONCLUSIONS: Hypernatremia was associated with poorer outcomes in patients with severe TBI. This finding warrants further investigation in a prospective, randomized study.
-
-
RESULTS: A total of 85,579 patients without a diagnosis of hypernatremia (control group) and 4542 patients with a diagnosis of hypernatremia (hypernatremia group) were identified. When controlling for age, comorbidities, gender, and cerebral edema, hypernatremia was associated with an increased rate of in-hospital mortality (odds ratio, 1.51; 95% confidence interval, 1.39e1.65), a longer mean length of stay (23.65 vs. 12.12 days; P < 0.001), an increased rate of nonroutine hospital discharge (odds ratio, 2.58; 95% confidence interval, 2.28e2.92), and greater mean total hospital cost ($227,112 vs. $112,507; P < 0.001). The patients with
-
Key words Hypernatremia - Nationwide inpatient sample - Traumatic brain injury -
Abbreviations and Acronyms DI: Diabetes insipidus EVD: External ventricular drain GCS: Glasgow Coma Scale ICP: Intracranial pressure LOS: Length of stay NIS: Nationwide Inpatient Sample
INTRODUCTION
S
evere traumatic brain injury (TBI) has the potential to disrupt the bloodebrain barrier and cause egress of fluid into the intracellular and extracellular spaces, resulting in cerebral edema and increased intracranial pressure (ICP).1 Induced hypernatremia is frequently used as a method of reducing acute elevations in ICP after TBI.2,3 Hypernatremia is a physiological condition in which the homeostatic water balance is disrupted, resulting in an increased concentration of serum sodium. An increase in serum sodium concentration is believed to cause a decrease in cell volume and cerebral blood volume, thereby decreasing cerebral edema. It also can modulate neuroinflammatory pathways, restore neuronal membrane potentials, and reduce blood viscosity.4,5 For these reasons, hypertonic saline is frequently given as a bolus to induce short periods of hyperosmolarity after TBI.6 However, hypernatremia can occur as a result of repeated dosing. Although many have endorsed the therapeutic effects of hyperosmolar therapy, others have shown that hypernatremia after severe TBI is independently associated with poor patient
PEG: Percutaneous endoscopic gastrostomy TBI: Traumatic brain injury Department of Neurosurgery, State University of New York Upstate Medical University, Syracuse, New York, USA To whom correspondence should be addressed: Haydn Hoffman, M.D. [E-mail: [email protected]] Citation: World Neurosurg. (2018). https://doi.org/10.1016/j.wneu.2018.07.089 Journal homepage: www.WORLDNEUROSURGERY.org Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2018 Elsevier Inc. All rights reserved.
WORLD NEUROSURGERY -: e1-e7, - 2018
www.WORLDNEUROSURGERY.org
e1
ORIGINAL ARTICLE HAYDN HOFFMAN ET AL.
HYPERNATREMIA AFTER TRAUMATIC BRAIN INJURY
prognosis and death.7 Hypernatremia has also been associated with poorer outcomes in patients with mild TBI.8 It remains controversial whether hypernatremia independently worsens the outcomes or is simply a surrogate marker for illness severity. For example, severe TBI can be associated with central diabetes insipidus (DI), a potentially confounding variable that causes hypernatremia and independently increases mortality.7 The use of iatrogenic hypernatremia after TBI remains controversial and its effects on patient prognosis continue to be debated. This is reflected in the most recent guidelines for severe TBI, which do not offer a specific recommendation regarding the use of hyperosmolar therapy.9 Thus, we sought to assess the effect of hypernatremia on the outcomes of patients with severe TBI using a large national healthcare database. METHODS The nationwide inpatient sample (NIS) was used to identify the patients for present study. The NIS is the largest publicly available all-payer healthcare database. It includes inpatient data obtained from hospital discharges in all states participating in the Healthcare Cost and Utilization Project. Approximately 1000 hospitals are sampled annually to include 8 million discharges for a 20% stratified sample of U.S. community hospitals. Discharge weights are provided to obtain national estimates of disease incidence. The data include 15 diagnoses and procedures per discharge, provided as International Classification of Diseases, Ninth Revision, Clinical Modification codes. Patients aged 18 years who had been discharged from 2002 to 2011 with a primary discharge diagnosis of TBI were identified. The Clinical Classifications Software code 233 was used to identify patients with TBI. The Clinical Classifications Software code 233 encompasses several International Classification of Diseases, Ninth Revision, Clinical Modification codes for TBI (800.00e801.9 and 803.1e804.9 codes, fracture of cranial vault, skull base, or facial bone with intracranial injury; 850.0e854.1 codes, concussion, cerebral contusion, subdural hematoma, epidural hematoma, other and unspecified traumatic intracranial hemorrhage). Patients with a primary diagnosis of V1552 (history of TBI), elective hospital admission, or a secondary diagnosis of DI (code 253.5) were excluded. Because the presenting Glasgow Coma Scale (GCS) score was not available, patients with severe TBI were identified as those undergoing external ventricular drain placement (EVD; code 02.2), ICP monitor placement (code 01.10), craniectomy or craniotomy (code 01.24), hematoma drainage (codes 01.31 or 01.39), or mechanical ventilation (codes 96.04, 96.70, 96.71, 96.72). Those who had not undergone any of these procedures were excluded. Patients with a diagnosis of hypernatremia (code 276.0) were included in the hypernatremia group and those without this diagnosis were included in the control group of the present study. We did not have access to data regarding the duration, degree, or etiology of hypernatremia. The primary outcome in the present study was in-hospital mortality, and the secondary outcomes included the length of stay (LOS), nonroutine hospital discharge, total hospital charges, tracheostomy (codes 31.1, 31.21, 31.29), and percutaneous endoscopic gastrostomy (PEG; code 43.11). With the exception of the last 2 outcomes, these items are provided as discrete data
e2
www.SCIENCEDIRECT.com
elements for each hospital discharge in the NIS. The NIS classifies discharge disposition as routine, transfer to a short-term hospital, transfer to a skilled nursing facility or an intermediate care facility, and home healthcare. These were dichotomized to routine or nonroutine discharges. Age, race (classified as white, black, Hispanic, Asian or Pacific Islander, Native American, or other), sex, and 29 Elixhauser comorbidities were adjusted for in the outcome analysis. The Elixhauser comorbidity measures cover a broad range of conditions that are independently associated with increased LOS, hospital charges, and mortality and are provided in the NIS data set.10 Univariate analyses using Student’s t test for continuous variables and Fisher’s exact test for categorical variables were performed to compare the demographic data, hospital characteristics, and comorbidities in patients with and without hypernatremia. A multiple logistic regression analysis was used to compare the incidences of potential predictors of hypernatremia in the hypernatremia and control groups. Multiple logistic regression or multiple linear regression analyses were performed to compare binomial or continuous outcomes between each group, respectively. Age, sex, race, and Elixhauser comorbidities were adjusted for in each regression analysis. The presence of cerebral edema (code 348.5) was adjusted for in the outcomes analysis. Statistical analysis was performed using IBM SPSS Statistics, version 24.0 (IBM Corp., Armonk, NY). RESULTS The total number of hospital discharges from 2002 to 2011 with a primary diagnosis of TBI was 331,759. After excluding cases of nonsevere TBI and those with a secondary diagnosis of DI, 90,121 cases were included in the present study. Of these, 10,830 (12.0%) required EVD or ICP monitor placement, 38,015 (42.2%) underwent craniotomy/craniectomy with or without hematoma evacuation, and 66,606 (73.9%) required mechanical ventilation. A total of 85,579 patients (95.0%) did not have a diagnosis of hypernatremia (control group) and 4542 (5.0%) did have hypernatremia (hypernatremia group). In the univariate analyses, the hypernatremia group had a slightly older mean age and included more men and more patients treated in urban and teaching hospitals. The full list of demographic data and comorbidities is included in Table 1. Patients aged 36e65 years had a greater incidence of hypernatremia compared with patients aged <36 years. Female sex was associated with a lower odds of hypernatremia than male sex. Compared with white patients, none of the other races had greater odds of developing hypernatremia. Coagulopathy and hypernatremia were significantly associated. The complete list of the predictors of hypernatremia is presented in Table 2. The outcomes were worse in the hypernatremia group than in the control group (Table 3 and Figure 1). A binary logistic regression analysis was performed to evaluate the association between hypernatremia and in-hospital mortality. The logistic regression model was highly statistically significant (c2 ¼ 6790.35; P < 0.001). The rate of in-hospital mortality was 24.8% in the control group and 30.7% in the hypernatremia group (odds ratio, 1.51; 95% confidence interval, 1.39e1.65; P < 0.001). A multiple regression analysis was run to predict the LOS from hypernatremia and the additional covariates. These variables significantly predicted the LOS [F(34,
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2018.07.089
ORIGINAL ARTICLE HAYDN HOFFMAN ET AL.
HYPERNATREMIA AFTER TRAUMATIC BRAIN INJURY
Table 1. Comparison of Demographic, Hospital Characteristics, and Comorbidities in Patients with Traumatic Brain Injury with and without Hypernatremia Hypernatremia (n, %) Variable Patients (n) Mean age (years)
No
Yes 4542
NA
54.59 21.66
0.001 <0.001
24,348 (28.5)
1117 (24.6)
36e65
30,581 (35.7)
1782 (39.2)
>65
30,650 (35.8)
1643 (36.2)
Male
60,972 (71.3)
3381 (74.5)
Female
24,508 (28.7)
1160 (25.5)
Race
Hypertension
Yes
1213 (1.4)
137 (3.0)
27,642 (32.7)
1620 (35.9)
Obesity
1464 (1.7)
131 (2.9)
Renal failure
3358 (4.0)
355 (7.9)
Mean number of chronic conditions
3.62 2.81
4.72 3.02
P Value
<0.001
NA, not applicable.
0.207
White
46,269 (69.1)
2615 (69.6)
Black
7243 (10.8)
426 (11.3)
Hispanic
8482 (12.7)
433 (11.5)
Asian/Pacific Islander
1843 (2.8)
94 (2.5)
Native American
575 (0.9)
38 (1.0)
Other
2541 (3.8)
153 (4.1)
Northeast
15,723 (18.4)
747 (16.4)
Midwest
17,428 (20.4)
958 (21.1)
South
33,761 (39.5)
1948 (42.9)
West
18,667 (21.8)
889 (19.6)
2927 (3.5)
148 (3.3)
<0.001
Hospital region
Hospital bed size
16,228 (19.2)
770 (17.3)
Large
65,533 (77.4)
3541 (79.4) <0.001
Location Urban
82,308 (97.2)
4377 (98.2)
Rural
2380 (2.8)
82 (1.8)
Teaching
62,579 (73.9)
3450 (77.4)
Nonteaching
22,109 (26.1)
1009 (22.6)
16,438 (19.5)
758 (16.8)
Congestive heart failure
5157 (6.1)
487 (10.8)
Chronic pulmonary disease
997 (1.2)
136 (0.2)
<0.001
Hospital teaching status
<0.001
Comorbidities
Continues
WORLD NEUROSURGERY -: e1-e7, - 2018
2780901) ¼ 327.20; P < 0.001]. The mean LOS was lower in the control group (12.12 days) than in the hypernatremia group (23.65 days; P < 0.001). A binary logistic regression analysis was also performed to evaluate the association between hypernatremia and nonroutine hospital discharge. The logistic regression model was highly statistically significant (c2 ¼ 12016.98; P < 0.001). The rate of nonroutine hospital discharge was 70.3% in the control group compared with 91% in the hypernatremia group (odds ratio, 2.58; 95% confidence interval, 2.28e2.92; P < 0.001). Hypernatremia was associated with a greater mean total hospital cost ($227,112) compared with the controls group ($112,507; P < 0.001). Both tracheostomy and PEG placement were performed more frequently in the hypernatremia group (Table 3). DISCUSSION
0.005
Medium
Alcohol abuse
Diabetes with chronic complications
No
<0.001
Sex
Small
P Value
85,579
18e35
Hypernatremia (n, %) Variable
53.44 22.71
Age group (years)
Table 1. Continued
The use of hypertonic saline to achieve hypernatremia in patients with TBI has been common practice in neurointensive care units. Noniatrogenic factors can also contribute to hypernatremia, such as hypovolemia, insensible free water losses, and DI. Hypernatremia is used therapeutically to decrease cerebral edema and ICP with the goal of reducing secondary injury.11 Compared with other therapies for intracranial hypertension such as hyperventilation and barbiturate coma, hypernatremia might be associated with fewer adverse effects. This has been challenged, however, with several studies finding a correlation between hypernatremia and worse clinical outcomes in patients with TBI.7,8,12-14 This seems plausible, because hypernatremia has been shown to be associated with worse outcomes in other intensive care unit settings.15 Given the frequency with which hypernatremia is applied in TBI, this therapy warrants close scrutiny. To the best of our knowledge, the present study is the first to use the NIS to evaluate the outcomes associated with hypernatremia after TBI. The univariate analysis revealed that the patients with hypernatremia were older and tended to have more comorbidities on admission to the hospital. Patients aged 36e65 years had a greater incidence of hypernatremia compared with those aged <36 years.
www.WORLDNEUROSURGERY.org
e3
ORIGINAL ARTICLE HAYDN HOFFMAN ET AL.
HYPERNATREMIA AFTER TRAUMATIC BRAIN INJURY
Table 2. Multivariate Analysis of Potential Predictors of Hypernatremia Development Variable
P Value
OR
95% CI
<0.001
1.19
1.09e1.30
0.074
0.91
0.82e1.01
Black
0.996
1.00
0.90e1.11
Hispanic
0.079
0.91
0.82e1.01
Asian or Pacific Islander
0.126
0.85
0.68e1.05
Native American
0.356
1.17
0.84e1.64
Other
0.567
1.05
0.89e1.25
Female sex
<0.001
0.80
0.74e0.87
Alcohol abuse
<0.001
0.78
0.71e0.86
Congestive heart failure
<0.001
1.70
1.51e1.91
Coagulopathy
<0.001
2.41
2.20e2.63
Diabetes
<0.001
1.61
1.31e1.97
Drug abuse
0.051
0.86
0.73e1.00
Hypertension
0.051
1.08
1.00e1.17
Obesity
0.003
1.37
1.12e1.68
Age (years) 36e65 >65 Race
Comorbidities
Renal failure Valvular heart disease
<0.001
1.45
1.26e1.67
0.075
0.84
0.70e1.02
CI, confidence interval; OR, odds ratio.
Furthermore, female sex was associated with a lower odds of having hypernatremia than the male sex. This is consistent with other studies illustrating a decline in total water body content in older patients and women, in particular, which could predispose them to hypernatremia.16 Similarly, of the comorbidities analyzed, we found that patients with coagulopathy and renal failure had some of the greatest odds of having hypernatremia. Increased rates of renal failure have been shown to be associated with hypernatremia,13 which might result from the kidney’s inability to maintain osmotic equilibrium. Coagulopathy was found to be significantly associated with hypernatremia. It is possible that patients treated with hypertonic fluids and osmotic diuretics were intravascularly depleted, making them more prone to consumption coagulopathies. Another possibility is that patients with hypernatremia had more severe TBI, which is known to be a risk factor for the development of coagulopathy.17 However, all the patients in present study had severe TBI and were at risk of developing an associated coagulopathy. The patients in the hypernatremia group had greater mortality rates and longer LOSs than those in the control group. The rates of tracheostomy, PEG placement, and nonroutine hospital discharge were greater in the hypernatremia group, suggesting poorer functional outcomes in those who did survive. Because
e4
www.SCIENCEDIRECT.com
severe TBI is associated with the development of central DI, this could have resulted in confounders in studies assessing the relationship between hypernatremia and outcomes.8,18,19 Thus, DI was a part of our exclusion criteria. It is possible that neurointensivists are more likely to aggressively treat severe TBI with hypernatremia rather than moderate or mild cases; thus, we excluded nonsevere injuries. The NIS does not directly specify the severity of TBI. Therefore, we selected for cases of severe TBI by identifying those who required an EVD or another form of ICP monitoring, because the guidelines for management of severe TBI recommend ICP monitoring in patients with a GCS of 8.9 This GCS threshold is also used as an indication for intubation20; thus, mechanical ventilation was added to the inclusion criteria. One could argue that the patients in the hypernatremia group were more likely to have experienced cerebral edema, which could have confounded the outcomes. However, we adjusted for the presence of cerebral edema in the analysis to account for this. Although our study is the first to use a large healthcare database to investigate this topic, it is not the first to identify worse outcomes in those with hypernatremia. In a retrospective analysis of patients with severe TBI from a single institution, Vedantam et al.8 demonstrated a greater mortality rate in those with hypernatremia. Both Li et al.12 and Aiyagari et al.13 showed that an increasing severity of hypernatremia was associated with an increasing incidence of mortality. Additionally, Shehata et al.,21 in a singleinstitution study, reported greater mortality and longer LOSs for TBI patients with hypernatremia compared with those who were normonatremic. Our mortality rate of 30.7% in the hypernatremia group is comparable to the rate of 26% reported by Maggiore et al.7 in the patients with hypernatremia from their cohort. Our rate is also similar to the overall mortality rate of 30% in hypernatremic patients reported by Aiyagari et al.13 but substantially lower than the incidence of 67% reported by Li et al.12 In the latter study, most (54%) of the patients had severe hypernatremia (>160 mEq/L), which might have contributed to their greater mortality rate. A number of reasons are possible for why hypernatremia could contribute to worse outcomes after TBI. Although hypernatremia can reduce cerebral edema, the resulting homeostatic imbalance might be deleterious toward other organ systems.22 Hypernatremia has been associated with greater rates of renal failure, which can occur through renal vasoconstriction, resulting in a decreased glomerular filtration rate.23,24 Accordingly, Froelich et al25 found that in neurocritically ill patients, continuous hypertonic saline therapy was associated with greater levels of creatinine and blood urea nitrogen. Hypernatremic states have also been shown to be associated with rhabdomyolysis.26 Hypernatremia could also have negative effects on cardiac contractility.27 Additionally, although the goal of induced hypernatremia is generally to avoid secondary brain injury, some studies have suggested it might actually contribute to this. In animal studies, hypernatremia has been shown to cause myelinolysis and cellular necrosis.28 Furthermore, attempts to correct hypernatremia too quickly can cause rebound elevations in ICP, cerebral edema, and seizures.29 The use of a nationwide data set gives our results strong generalizability. Most of the reported data evaluating the effect of hypernatremia on the outcomes after TBI have come from singleinstitution analyses. However, inherent limitations exist to the use
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2018.07.089
ORIGINAL ARTICLE HAYDN HOFFMAN ET AL.
HYPERNATREMIA AFTER TRAUMATIC BRAIN INJURY
Table 3. Outcomes of Patients with and without Hypernatremia Hypernatremia (n, %) Outcome
No
Yes
P Value
OR
95% CI
Mortality
21,200 (24.8)
1394 (30.7)
<0.001
1.51
1.39e1.65
12.12 15.71
23.65 22.74
<0.001
NA
NA
Mean LOS (days) Nonroutine disposition Mean total hospital cost ($)
60,162 (70.3)
4134 (91.0)
<0.001
2.58
2.28e2.92
112,507 138,145
227,112 196,669
<0.001
NA
NA
Tracheostomy
12,589 (14.7)
1777 (39.1)
<0.001
2.23
2.04e2.42
PEG placement
9566 (11.2)
1401 (30.8)
<0.001
1.95
1.78e2.13
CI, confidence interval; LOS, length of stay; NA, not applicable; OR, odds ratio; PEG, percutaneous endoscopic gastrostomy.
of the NIS that should be considered when interpreting these results. We did not have access to each patient’s presenting GCS; thus, the presence of severe TBI was determined by surrogate factors, such as the need for ICP monitoring or mechanical
ventilation. Given that the incidence of mechanical ventilation in patients with severe TBI nears 100% in reported cohorts,7,30-32 we believe we identified the vast majority of patients with severe TBI with data in the NIS database. The sodium level for diagnosing
Figure 1. Primary and secondary outcomes in the hypernatremia group compared with the control group (**P < 0.001).
WORLD NEUROSURGERY -: e1-e7, - 2018
www.WORLDNEUROSURGERY.org
e5
ORIGINAL ARTICLE HAYDN HOFFMAN ET AL.
HYPERNATREMIA AFTER TRAUMATIC BRAIN INJURY
hypernatremia could have varied between institutions, and the duration and etiology of hypernatremia for each patient was unknown. Furthermore, the use of a validated functional outcome measurement in TBI such as the Glasgow Outcome Scale would have been preferred. Instead, we had to use surrogate outcome measures such as tracheostomy or PEG placement and discharge disposition. It is possible that patients with nonroutine dispositions at hospital discharge eventually improved enough to regain functional independence. Also, the retrospective nature prevented us determining causation. Likewise, the controversy regarding whether hypernatremia is merely a surrogate for illness severity could not be addressed in the present study. We only included patients with severe TBI to minimize the confounding effects of injury severity. Additional information about each patient’s injury, such as the location of the lesions and the amount of cerebral
REFERENCES 1. Tan SK, Kolmodin L, Sekhon MS, Qiao L, Zou J, Henderson WR, et al. The effect of continuous hypertonic saline infusion and hypernatremia on mortality in patients with severe traumatic brain injury: a retrospective cohort study. Can J Anaesth. 2016;63:664-673. 2. Peterson B, Khanna S, Fisher B, Marshall L. Prolonged hypernatremia controls elevated intracranial pressure in head-injured pediatric patients. Crit Care Med. 2000;28:1136-1143. 3. Mangat HS, Chiu YL, Gerber LM, Alimi M, Ghajar J, Härtl R. Hypertonic saline reduces cumulative and daily intracranial pressure burdens after severe traumatic brain injury. J Neurosurg. 2015;122:202-210.
edema present, would have aided in matching the groups; however, these data are not provided in the NIS database. Despite these limitations, we identified a robust association between hypernatremia and a variety of unfavorable outcomes that merits close scrutiny of this therapy. Evaluation in a prospective, randomized study is warranted to assess for causality. Such a study will need to control for variations in the amount of cerebral edema, the severity of ICP elevation, and the number of hyperosmolar therapies administered. CONCLUSIONS After severe TBI, hypernatremia has been associated with a greater incidence of in-hospital mortality, longer LOS, and nonroutine discharge disposition. A prospective, randomized study on the effects of hypernatremia after TBI is merited.
management of severe traumatic brain injury, fourth edition. Neurosurgery. 2017;80:6-15.
patients with serious head injury. BMJ. 1993;307: 547-552.
10. Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care. 1998;36:8-27.
21. Shehata M, Ragab D, Khaled M, Hegazy M, Hussein A, Khaled H. Impact of hypernatremia on patients with sever traumatic brain injury. Crit Care. 2010;14:355.
11. Khanna S, Davis D, Peterson B, Fisher B, Tung H, O’Quigley J, et al. Use of hypertonic saline in the treatment of severe refractory posttraumatic intracranial hypertension in pediatric traumatic brain injury. Crit Care Med. 2000;28:1144-1151. 12. Li M, Hu YH, Chen G. Hypernatremia severity and the risk of death after traumatic brain injury. Injury. 2013;44:1213-1218. 13. Aiyagari V, Deibert E, Diringer MN. Hypernatremia in the neurologic intensive care unit: how high is too high. J Crit Care. 2006;21:163-172.
4. Murphy N, Auzinger G, Bernel W, Wendon J. The effect of hypertonic sodium chloride on intracranial pressure in patients with acute liver failure. Hepatology. 2004;39:464-470.
14. Kolmodin L, Sekhon MS, Henderson WR, Turgeon AF, Griesdale DE. Hypernatremia in patients with severe traumatic brain injury: a systematic review. Ann Intensive Care. 2013;3:35.
5. Muizelaar JP, Wei EP, Kontos HA, Becker DP. Cerebral blood flow is regulated by changes in blood pressure and in blood viscosity alike. Stroke. 1986;17:44-48.
15. Waite MD, Fuhrman SA, Badawi O, Zuckerman IH, Franey CS. Intensive care unitacquired hypernatremia is an independent predictor of increased mortality and length of stay. J Crit Care. 2013;28:405-412.
6. Kahle KT, Walcott BP, Simard JM. Continuous hyperosmolar therapy for traumatic brain injuryassociated cerebral edema: as good as it gets, or an iatrogenic secondary insult? J Clin Neurosci. 2013;20:30-31. 7. Maggiore U, Picetti E, Antonucci E, Parenti E, Regolisti G, Mergoni M, et al. The relation between the incidence of hypernatremia and mortality in patients with severe traumatic brain injury. Crit Care. 2009;13:R110. 8. Vedantam A, Robertson CS, Gopinath SP. Morbidity and mortality associated with hypernatremia in patients with severe traumatic brain injury. Neurosurg Focus. 2017;43:E2. 9. Carney N, Totten AM, O’Reilly C, Ullman JS, Hawryluk GW, Bell MJ, et al. Guidelines for the
e6
www.SCIENCEDIRECT.com
16. Kugler JP, Hustead T. Hyponatremia and hypernatremia in the elderly. Am Fam Physician. 2000;15: 3623-3630. 17. Laroche M, Kutcher ME, Huang MC, Cohen MJ, Manley GT. Coagulopathy after traumatic brain injury. Neurosurgery. 2012;70:1334-1345. 18. Hadjizacharia P, Beale EO, Inaba K, Chan LS, Demetriades D. Acute diabetes insipidus in severe head injury: a prospective study. J Am Coll Surg. 2008;207:477-484. 19. Boughey JC, Yost MJ, Bynoe RP. Diabetes insipidus in the head-injured patient. Am Surg. 2004;70: 500-503. 20. Gentleman D, Dearden M, Midgley S, Maclean D. Guidelines for resuscitation and transfer of
22. Lindner G, Funk GC. Hypernatremia in critically ill patients. J Crit Care. 2013;28:216. 23. Wilcox CS. Regulation of renal blood flow by plasma chloride. J Clin Invest. 1983;71:726-735. 24. Gerber JG, Branch RA, Nies AS, Hollifield JW, Gerkens JF. Influence of hypertonic saline on canine renal blood flow and renin release. Am J Physiol. 1979;237:F441-F446. 25. Froelich M, Ni Q, Wess C, Ougorets I, Härtl R. Continuous hypertonic saline therapy and the occurrence of complications in neurocritically ill patients. Crit Car Med. 2009;37:1433-1441. 26. Kung AW, Pun KK, Lam KS, Yeung RT. Rhabdomyolysis associated with cranial diabetes insipidus. Postgrad Med J. 1991;67:912-913. 27. Kozeny GA, Murdock DK, Euler DE, Hano JE, Scanlon PJ, Bansal VK, et al. In vivo effects of acute changes in osmolality and sodium concentration on myocardial contractility. Am Heart J. 1985;109:290-296. 28. Ayus JC, Armstrong DL, Arieff AI, Gabbai R, ID G. Effects of hypernatraemia in the central nervous system and its therapy in rats and rabbits. J Physiol. 1996;492:243-255. 29. De Petris L, Luchetti A, Emma F. Cell volume regulation and transport mechanisms across the blood-brain barrier: implications for the management of hypernatraemic states. Eur J Pediatr. 2001;160:71-77. 30. Khajavikhan J, Aminolah V, Khani A, Jaafarpour M, Kokhazade T. Outcome and predicting factor following severe traumatic brain injury: a retrospective cross-sectional study. J Clin Diagn Res. 2016;10:PC16-PC19.
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2018.07.089
ORIGINAL ARTICLE HAYDN HOFFMAN ET AL.
31. Godbolt AK, Stenberg M, Jakobsson J, Sorjonen K, Krakau K, Stålnacke BM, et al. Subacute complications during recovery from severe traumatic brain injury: frequency and associations with outcome. BMJ Open. 2015;5. e007208. 32. Haddad S, Aldawood AS, Alferayan A, Russell NA, Tamim HM, Arabi YM. Relationship between intracranial pressure monitoring and outcomes in
HYPERNATREMIA AFTER TRAUMATIC BRAIN INJURY
severe traumatic brain injury patients. Anaesth Intensive Care. 2011;39:1043-1050.
Citation: World Neurosurg. (2018). https://doi.org/10.1016/j.wneu.2018.07.089 Journal homepage: www.WORLDNEUROSURGERY.org
Conflict of interest statement: The authors declare that the article content was composed in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2018 Elsevier Inc. All rights reserved.
Received 12 April 2018; accepted 10 July 2018
WORLD NEUROSURGERY -: e1-e7, - 2018
www.WORLDNEUROSURGERY.org
e7