Perioperative estimation of the intracranial pressure using the optic nerve sheath diameter during liver transplantation


  • This study was funded by departmental sources.

Address reprint requests to Vijay Krishnamoorthy, M.D., Department of Anesthesiology, Harborview Medical Center, University of Washington, 325 Ninth Avenue, Box 359724, Seattle, WA 98104. Telephone: 206-744-4371; FAX: 206-744-8090; E-mail:


An elevation of the intracranial pressure (ICP) secondary to cerebral edema is a major contributor to morbidity and mortality in acute liver failure. In addition, invasive ICP monitoring in this setting is controversial because coagulopathy predisposes patients to hemorrhagic complications. In this case report, we describe the novel use of optic nerve sheath diameter monitoring as a noninvasive modality for checking for acute elevations in ICP in this setting. Because of the merits of rapidly evolving ultrasound technologies, this may serve as a safe method for improving patient care in this setting. Liver Transpl 19:246–249, 2013. © 2013 AASLD.


intracranial pressure


optic nerve sheath diameter.

Intracranial hypertension secondary to cerebral edema may contribute to mortality in 50% to 80% of patients with acute liver failure.[1] Although invasive intracranial pressure (ICP) monitoring is mentioned as an option by the Acute Liver Failure Study Group in its recommendations,[2] the benefits of ICP monitoring are unproven in this setting. ICP monitoring has been reported in the setting of fulminant hepatic failure for patients with grade III or IV encephalopathy and for patients undergoing liver transplantation.[3] However, the clinical use of ICP monitoring via traditional modalities in patients with acute liver failure is limited by associated coagulopathy, particularly during liver transplantation and the perioperative period, during which ICP monitor placement is associated with a 10% to 20% potential for the development of hemorrhagic complications.[2]

Despite the technical challenges associated with invasive ICP monitoring in patients with hepatic dysfunction, perioperative ICP monitoring may be beneficial for guiding the administration of targeted therapy (ie, mannitol, hypertonic saline, hyperventilation, and reverse Trendelenburg positioning) to prevent brain herniation in patients with cerebral edema due to hepatic dysfunction. The measurement of the optic nerve sheath diameter (ONSD) via optic nerve ultrasound has been shown to noninvasively and reliably detect elevations in ICP in a diverse cohort of neurocritical care patients in multiple studies.[4-6] A cutoff > 0.48 cm has been associated with ICP values > 20 mm Hg.[4] However, the use of this technique has never been reported for perioperative ICP monitoring in patients with hepatic dysfunction. Here we describe our experience with using the ONSD for perioperative ICP monitoring to guide therapy in a patient undergoing liver transplantation.


A previously healthy 40-year-old female without a past medical history presented to our hospital with progressive jaundice, dark urine, and bilateral lower extremity edema for 2 weeks. She denied a prior history of liver disease, exposure to toxins, alcohol use, recent viral illness, or recent international travel. Her vital signs at admission included a temperature of 37.6 °C, a heart rate of 96 beats per minute, and a blood pressure of 124/62 mm Hg. Her physical examination was remarkable for scleral icterus, bilateral lower extremity pitting edema, mild hepatomegaly, and a normal mental status. There was no evidence of cutaneous bruising, and her cardiac and pulmonary examinations were unremarkable. Her laboratory parameters at admission were a prothrombin time of 44.2 seconds, an international normalized ratio of 4.7, a total bilirubin level of 14.0 mg/dL, an aspartate aminotransferase level of 1402 U/L, an alanine aminotransferase level of 806 U/L, a hemoglobin level of 7.7 g/dL, a platelet count of 387,000 μl, and a serum creatinine level of 1.0 mg/dL. An abdominal computed tomography scan demonstrated an enlarged liver without ascites. The patient was admitted to our hospital for workup and supportive therapy for presumed acute liver failure of an unknown etiology.

Over the course of the next week, the patient's liver function deteriorated, and the patient developed acute kidney injury (with anuria and a creatinine level rising to 5.4 mg/dL) requiring dialysis catheter placement. Liver biopsy demonstrated severe hepatocellular necrosis. On hospital day 6, the patient experienced increasing somnolence and was unable to follow commands. She was eventually unable to protect her airway or manage her secretions, and this necessitated tracheal intubation and mechanical ventilation. At this time, a head computed tomography scan demonstrated effacement of the cortical sulci and basal cisterns and suggested generalized cerebral edema. The intensive care team considered invasive ICP monitoring but decided against it because of her coagulopathy. On hospital day 6, the patient was listed as status 1 for liver transplantation because of her deteriorating status, worsening encephalopathy, and multiorgan failure.

Among the issues facing this patient, one concern for the perioperative care team was optimal monitoring and management of her cerebral edema during general anesthesia. Clinically, the patient was not moving her extremities in response to commands but was able to withdraw in response to deep stimulation; she occasionally had spontaneous eye opening, and she had preserved brainstem reflexes (cough, corneal, and gag). Because invasive ICP monitoring was contraindicated by her coagulopathy, starting on hospital day 6 and before surgery, we elected to further monitor the patient's ONSD by ultrasound. All scans were performed by the same operator (V.K.), who had formal training in general critical care ultrasonography, with the same technique previously described in the literature.[4] The baseline preoperative optic nerve ultrasound examination was performed in the intensive care unit for a baseline ONSD value, which was 0.35 cm (Fig. 1); a value of 0.48 cm was chosen as the cutoff for an ICP > 20 mm Hg on the basis of the largest study available in the literature.[4] For measuring the ONSD, a linear ultrasound probe (SonoSite, Bothell, WA) with a standard vascular preset at a frequency of 7 MHz was placed on the superior and lateral margin of the orbit on the closed eyelid; an image of the globe with the retina and the optic nerve was obtained, and the diameter of the optic nerve sheath was identified 3 mm below the retina. This examination was repeated 3 times above each eye, and the average of the values was obtained. A representative image is shown in Fig. 1.

Figure 1.

Preoperative optic nerve sheath ultrasound demonstrating an ONSD of 0.35 cm.

On the following day (hospital day 7), our team was notified that a suitable donor was available for orthotopic liver transplantation. The patient was taken to the operating room, and standard American Society of Anesthesiologists monitors were set. The induction of general anesthesia was performed via the patient's preexisting endotracheal tube with isoflurane and intravenous fentanyl. The maintenance of general anesthesia for the 6-hour procedure was achieved via a balanced anesthetic technique with isoflurane, fentanyl, and vecuronium. Because of anticipated fluid and electrolyte shifts (including blood loss and resuscitation with crystalloids, colloids, and blood products) that could exacerbate her cerebral edema, we monitored the ONSD at 1-hour intervals during the surgical procedure. During the anhepatic phase (hour 4), an increase in the optic nerve diameter to 0.46 cm was observed (Fig. 2). Because of the increasing ONSD and our concern about an impending high ICP, we instituted increased reverse Trendelenburg positioning. Over the ensuing 30 minutes, the liver graft was reperfused, and the optic nerve diameter decreased to 0.39 cm over the next 3 hours. The remainder of the surgical procedure was uneventful, and the patient was transferred to the intensive care unit, at which, 4 hours after the operation, the optic nerve diameter was at the baseline level. The patient's trachea was extubated on the fourth postoperative day. The patient's neurological status and renal function improved, and the patient was discharged to a rehabilitation facility without any neurological deficits and without the need for dialysis.

Figure 2.

Increase in the ONSD to 0.46 cm at hour 4 during the transplant.


The use of ultrasound to measure the optic nerve diameter has been described in various papers and case series for more than a decade, so there is extensive literature support for this technique.[4-6] Traditionally, ICP has been measured invasively with either an intraparenchymal probe or an intraventricular catheter. Both techniques carry the risk of infection or bleeding[4, 7] even in the noncoagulopathic patient population. The noninvasive monitoring of an elevated ICP with optic nerve ultrasonography can, therefore, be advantageous in patients with signs of an increased ICP in the setting of acute fulminant liver failure. In fact, this examination has been described in the pediatric liver failure population; in a study assessing 22 pediatric patients with acute liver failure who were awaiting liver transplantation, optic nerve ultrasound correctly identified patients with a poor prognosis secondary to a raised ICP and was able to correctly risk-stratify survivors from nonsurvivors.[3] Our report is the first to describe the use of the ONSD for monitoring cerebral edema during general anesthesia for orthotopic liver transplantation in an adult patient.

As previously described,[4, 8] we identified the diameter of the optic nerve sheath 3 mm posterior to the retina and obtained all of our measurements at this distance. In this case, the sequential increase in the ONSD during surgery led us to institute a treatment to prevent a critical increase in ICP because published guidelines recommend keeping it at ≤20 mm Hg in the setting of traumatic brain injury and other forms of acute brain injury.[9, 10] The underlying mechanism of optic nerve sheath distension is likely the transmission of a raised ICP through the subarachnoid space to the optic nerve sheath; this is similar to the mechanism of papilledema. Unlike papilledema, however, optic nerve sheath distension occurs rapidly (within seconds of an acute rise in ICP).[4, 11, 12] This makes optic nerve ultrasound quite useful for the detection of acute elevations in ICP.

Recent studies have shown and validated that an increase in the ONSD anywhere from 4.5 to 5.5 mm[6, 13, 14] is associated with an ICP > 20 mm Hg. To date, Rajajee et al.[4] have performed the largest study correlating the ONSD with simultaneous invasive ICP monitoring: they correlated a total of 536 ONSD measurements to invasive ICP measurements in a heterogeneous group of 65 patients with acute brain injury, including both hemorrhagic and ischemic insults. They found that the optimal ONSD for the detection of an acutely increased ICP > 20 mm Hg was greater than 0.48 cm with a sensitivity of 96% and a specificity of 94%.[4]

The high negative predictive value reported in the literature for optic nerve ultrasound may be useful for excluding an elevated ICP in the complex intensive care unit patient with acute liver failure and cerebral edema for whom significant risks are associated with invasive ICP monitoring secondary to coagulopathy. In our specific patient scenario, optic nerve ultrasound proved to be a valuable monitoring tool in a patient who otherwise would have had no ICP monitoring; further studies are most certainly needed to assess whether perioperative ICP monitoring with optic nerve ultrasound will alter management and improve outcomes. In our particular patient, a successful liver transplant outcome was possible with the aid of this technique. In addition, because of the noninvasive nature of ultrasound technology and the minimal examination time for acquiring accurate measurements, the safety of optic nerve ultrasound in comparison with other more invasive modalities is maximized, with no reports of adverse events from the brief use of ultrasound for measuring ONSD.

Although this is promising, several limitations to the acquisition of reliable ONSD data and management do exist: although the largest study used a cutoff of 0.48 cm for predicting an ICP > 20 mm Hg with good sensitivity and specificity, several studies have used different cutoff values, with some as high as 0.70 cm.[5] The majority of the studies have been performed by clinicians with experience in the technique; a clinician new to this technique may require time to adequately acquire reliable and reproducible images while minimizing the inclusion of artifacts such as shadowing. In addition, the quality of the ultrasound probe itself may lead to higher interobserver variability, with older probes fairing more poorly.[12] Finally, although this is excellent for steady changes in ICP, it may not be the best monitor for acute fluctuations in ICP[15]; thus, it may aid in stratifying patients (especially intubated/sedated patients with several comorbidities) who require invasive monitoring, which may be more appropriate for the management of acute changes in ICP.

The advent of ultrasound technology is enhancing the practice of anesthesiology from transesophageal echocardiography to regional anesthesia. The use of these technologies has the potential to improve the safety and quality of the care that we can deliver to our sickest patients. Although large studies are needed to verify the utility of ONSD monitoring in the setting of acute liver failure, we hope that our case report can begin to open the door to beginning those studies.