Cerebral edema with accompanying elevation in intracranial pressure may complicate cardiac arrest and subsequent management in the intensive care unit (ICU), especially in patients requiring renal replacement therapy. Ensuring stability of serum osmolality and sodium is particularly challenging during renal replacement therapy.1 Commercial preparations of continuous renal replacement therapy (CRRT) therapy fluid contain 140 mEq/L sodium and represent hypotonic solution when corrected for serum protein content after infusion.1 The use of hypertonic saline solution (HTS) in conjunction to renal replacement therapy can effectively stabilize serum osmolality and prevent cerebral edema after ischemic injury.
In late October of 2018, a previously healthy 22 year old male collapsed with cardiac arrest during a soccer match. Attempts at resuscitation by bystanders and several rounds of defibrillation for ventricular fibrillation lasted 75 minutes before he regained perfusing sinus rhythm. Upon arriving to the emergency department (ED), he was administered intravenous (IV) norepinephrine, vasopressin, and epinephrine infusions for blood pressure support. He was also intubated and mechanically ventilated with additional hypothermia protocol for central nervous system protection. Of note, he received several doses of hypertonic NaHCO3 solutions (8.4%, 50 ml each) in the ED and one dose of IV hydrocortisone 100 mg shortly after arrival to the ICU. However, even with maximal medical support, his cardiac index by Fick method was only 1.17 (reference range: 2.5–4). His O2 saturation on maximal ventilatory support and epoprostenol infusion was only at 72%. Therefore, veno-arterial extracorporeal membrane oxygenation (VA ECMO) was initiated in attempt to improve both cardiac output and oxygenation, with arterial cannula placed via right femoral artery. Baseline head computerized tomography (CT) scan was unremarkable for acute intracranial process and cardiovascular imaging procedures with two-dimensional echocardiogram and contrasted CT angiogram described global hypokinesis in addition to suspicion for thrombus formation in the left ventricle of the heart and aorta. Troponin peaking at 232 ng/ml (reference range: 0.0–0.04 ng/ml) suggested diffuse cardiac injury (Table 1). His serum creatinine was already elevated on arrival (1.6 mg/dl) and he displayed rising serum creatinine kinase with full anuria. He had ongoing and severe mixed acidosis (pH: 7, pCO2: 74 mm Hg, pO2: 58 mm Hg) and persistent lactate accumulation (7.16 mmol/L) as well (Table 1). Due to a clinical picture involving critical hemodynamic instability, rhabdomyolysis, acidosis, and progressive fluid accumulation with anuria, CRRT was initiated by the second hospital day, approximately 16 hours after hospital arrival.
Table 1. -
Key Electrolyte and Biochemistry During Admission and Hospitalization
|PaO2 (mm Hg)
|PaCO2 (mm Hg)
|Serum creatinine (mg/dl)
|Calcium (mg/dl) (uncorrected)
|White blood cell count (k/mm3)
|Red blood cell count (m/mm3)
|Lactic acid, plasma (mmol/L)
The first laboratory result collected on admission is baseline. Later laboratory results collected on the same day is day 0. When more than one laboratory result is available on a day, ranges are provided.
n/a, not available; PaO2, partial arterial pressure of O2; PaCO2, partial arterial pressure of CO2.
We utilized a third-generation, commercially available CRRT platform (Prismaflex; Baxter Healthcare Corp, Deerfield, IL) to deliver CRRT in hemodiafiltration functional modality and attached it in series with ECMO. A separate IV access was used to deliver the added HTS according to published protocol, with a goal to stabilize serum osmolality by stabilizing serum sodium at or slightly above the pre-CRRT range (target range of 148–150 mEq/L).1 Before starting CRRT, patient’s serum sodium was 149 mEq/L (Table 2). Therapy fluid was titrated to a global effluent rate of 50 ml/kg/h (4.3 L/h), with peripheral HTS 3% infusion at 110 ml/h (Table 2). Of note, the effluent fluid produced during the continuous veno-venous hemodiafiltration (CVVHDF) process did not show the expected pink discoloration of significant rhabdomyolysis. Within 6 hours after the commencement of CVVHDF, the patient’s hemodynamic state improved as he was able to be weaned off vasopressin and norepinephrine with norepinephrine requirements dropping to 4–6 µg/min.
Table 2. -
Serum Sodium and Osmolality
Before and During Treatment with CRRT
||Day After Starting CRRT + HTS
|Measured osmolality (mOsmol/kg)
|Calculated osmolality (mOsmol/kg)
|3% NaCl rate (ml/h)
|CVVHDF rate (L/h)
Admission is day 0. Day 1 is the first day with CRRT with HTS.
CRRT, continuous renal replacement therapy; CVVHDF, continuous veno-venous hemodiafiltration; HTS, hypertonic saline solution.
However, on the fourth day, the patient’s hemodynamics worsened with progressive norepinephrine requirements, rising CK and emerging ischemic demarcation of the right lower extremity, the site distal to VA ECMO cannulation. Extracorporeal membrane oxygenation output remained unchanged and there were no technical problems or clotting in either ECMO or CRRT circuits. Major differential at that point in time included impaired perfusion to the right lower extremity solely due to ECMO cannula alone or increasing vasodilation due to the profound state of inflammation induced by initial ischemic injury and not controlled by CRRT therapy alone. After careful deliberation we also advocated for added glucocorticoid therapy with hydrocortisone 100 mg IV every 8 hours.2 In addition, the surgical team proceeded with bedside cryoamputation of the right lower extremity with dry ice (crystallized CO2, temperature-79°C) at a level above the knee. Serum myoglobin level was markedly elevated at 143,030 ng/ml (reference range: 28–72 ng/ml) shortly after the initiation of cryoamputation. Hemodynamic instability and vasoactive medication requirement improved again after the addition of glucocorticoid therapy within 24 hours. Norepinephrine dose decreased to 12 µg/min from 35 µg/min. Despite cryoamputation and ongoing high-volume CRRT, the patient’s CK continued to rise and remained elevated for the next several days, peaking at 168,632 U/L (reference range: 30–260 U/L) (Figure 1A).
Throughout this period, therapy with CVVHDF helped our patient regain metabolic stability and serum sodium was maintained in target range for approximately 96 hours (Table 2) after which he was weaned off both ECMO and CRRT. Hypertonic saline solution was discontinued on day 5. Patient showed signs of stable cognitive functioning with ability to respond to commands as early as 3 days post-arrest. Brain magnetic resonance imaging (MRI) at 22 days post-arrest showed signs consistent with hypoxic brain injury (Figure 2). He was transferred to a hospital closer to his home for further rehabilitation 31 days post-arrest with new baseline creatinine at 1.9 mg/dl without renal replacement therapy requirements and ongoing neurologic recovery. Patient was subsequently lost to local clinical follow-up.
While the initial CT scan of our patient did not show brain edema, serum sodium was elevated at 149 mEq/L on presentation in the ICU due to the hypertonic NaHCO3 therapy during resuscitation. Brain edema is expected to peak at 48–72 hours after the initial injury. Hence, initiating CRRT with “standard” (140 mEq/L) solution may have exposed him to a further drop in serum sodium and a true potential for brain edema. While coexisting regional citrate anticoagulation would promote a relative hypernatremia, such tendency is contingent on the relative rate of flow of both the hemofiltration/dialysis fluids and of the citrate (acid-citrate dextrose A [ACD-A]) solution.1,3 In our index case, ACD-A use was not warranted due to ongoing systemic heparinization already ongoing for the VA ECMO therapy. We elected to deliver HTS, which contains 1,026 mOsm/L sodium1 to prevent the drop in serum sodium and osmolality during CRRT, achieving a measured rise of overall serum tonicity. Moreover, while we did not directly measure intracranial pressure because the patient’s baseline CT did not suggest cerebral edema or herniation, the MRI scan at 22 days post-arrest suggested that the patient sustained hypoxic ischemic brain injury (HIBI) sometime during his illness (Figure 2). As reviewed by Sekhon et al.,4 studies have shown that during HIBI, there is decreased oxygen delivery to neurons which die within minutes after the primary injury. Reinstatement of oxygen delivery in a metabolically active organ like the brain causes reperfusion injury which results in neuronal dysfunction and disruption of the blood brain barrier.4 This induces an increase in expression of aquaporin-4 and subsequent influx of water into the brain resulting in cerebral edema.4 Hypertonic saline solution has been shown to reduce blood brain barrier disruption mediated by aquaporin-4 channels in the astrocyte end-feet, and thus attenuates cerebral edema.5 Therefore, although we cannot confirm that our patient had increased intracranial pressure or cerebral edema, we can infer that the hypoxic process involved some degree of cerebral edema that was mitigated by stabilizing serum sodium and osmolality during the period of critical illness and ongoing CRRT therapy.
So far, there are few publications that document the use of HTS in addition to standard CRRT replacement fluid in the setting of critical illness. Rifkin et al.6 endorses the use of HTS with CVVHDF in a patient presenting with end-stage renal failure and cerebral edema due to increased intracranial pressure after routine hemodialysis. Their patient did not experience adverse effects during therapy and Rifkin et al.6 were able to reduce intracranial pressure but the patient’s neurologic status did not improve. Fülöp et al.1 also describe two cases where both patients sustained traumatic brain injury and HTS was used with CRRT. By setting the target serum sodium level between 150 and 155 mEq/L, intracranial pressure in both patients decreased. One patient ultimately experienced a near-complete recovery of his neurologic function and renal function; the other expired due to respiratory status deterioration.1 There is even an account where a young patient presenting after resuscitation from cardiac arrest experienced intracranial hypertension after starting CRRT without HTS.7 The patient was in a minimally conscious state 51 days post-arrest; at 3 months follow-up, the patient remained in a minimally conscious state with signs of slow improvement even with neurorehabilitation.7
Interestingly in our case, CK continued to rise despite cryoamputation of the right lower extremity (Figure 1A). Such continued rise of CK could have been attributable perhaps less to the right lower extremity ischemia alone but to diffuse muscle injury after the initial mechanical trauma and ischemic insult during prolonged cardiopulmonary resuscitation, further aggravated by the state of immobilization and paralysis in the ICU setting. The presence of AKI in published studies so far appears to identify a subcohort of the critically ill who may potentially benefit from an addition of stress dose glucocorticoid therapy.2
The complex mechanisms that involve cerebral injury in the face of hypoxic events and kidney injury have not been extensively studied. With progressive fluid overload and critical illness, serum albumin and total protein concentration are expected to drop. Therefore, the net tonicity of extracellular fluid—assuming identical measured serum sodium concentration—is expected to drop. Current literature is primarily focused on the changes of serum sodium, rather than total osmolality. While blood urea nitrogen is perceived not to be contributing to effective osmolality, such notion may be grossly inaccurate in those with severe AKI or end-stage renal failure receiving conventional, intermittent renal replacement therapy.8 Our case demonstrates the value of therapeutic maneuvers through renal replacement therapy. By controlling shifts in osmolarity and plasma tonicity, we may offer some benefit to the critically ill.
Administration of HTS with CRRT at aiming to maintain serum sodium levels in the therapeutic range of 148–150 mEq/L in a patient with hemodynamic instability may mitigate the detrimental effects of hypoxic brain injury. The use of HTS needs additional studies to explore the potential for reducing cerebral edema and improving outcomes in these patients in larger series and in controlled studies.
1. Fülöp T, Zsom L, Rodríguez RD, Chabrier-Rosello JO, Hamrahian M, Koch CA. Therapeutic hypernatremia management during continuous renal replacement therapy with elevated intracranial pressures and respiratory failure. Rev Endocr Metab Disord 2019; 20:65–75.
2. Annane D, Renault A, Brun-Buisson C, et al.; CRICS-TRIGGERSEP Network: Hydrocortisone plus fludrocortisone for adults with septic shock. N Engl J Med 2018; 378:809–818.
3. Fülöp T, Abdul Salim S, Zsom L. Regional citrate anticoagulation for continuous renal replacement therapy without post-filter monitoring of ionized calcium. J Renal Inj Prev 2018; 7:139–143.
4. Sekhon MS, Ainslie PN, Griesdale DE. Clinical pathophysiology of hypoxic ischemic brain injury after cardiac arrest
: a “two-hit” model. Crit Care 2017; 21:90.
5. Nakayama S, Migliati E, Amiry-Moghaddam M, Ottersen OP, Bhardwaj A. Osmotherapy with hypertonic saline attenuates global cerebral edema following experimental cardiac arrest
via perivascular pool of aquaporin-4. Crit Care Med 2016; 44:e702–e710.
6. Rifkin SI, Malek AR, Behrouz R. Use of hypertonic continuous venovenous hemodiafiltration to control intracranial hypertension in an end-stage renal disease patient. Int J Nephrol 2010; 2010:391656.
7. Lund A, Damholt MB, Strange DG, Kelsen J, Møller-Sørensen H, Møller K. Increased intracranial pressure during hemodialysis in a patient with anoxic brain injury. Case Rep Crit Care 2017; 2017:5378928.
8. Venkatasubba Rao CP, Bershad EM, Calvillo E, et al. Real-time noninvasive monitoring of intracranial fluid shifts during dialysis using Volumetric Integral Phase-Shift Spectroscopy (VIPS): A Proof-of-Concept Study. Neurocrit Care 2018; 28:117–126.