Noninvasive estimation of intracranial pressure with optic nerve sheath diameter sonography

Abstract

Background: Modern neurointensive care aims to prevent or mitigate secondary cerebral insults after severe primary cerebral insults, such as subarachnoid hemorrhage (SAH) or traumatic brain injury (TBI). Optimization of cerebral perfusion is central to this task. Elevated intracranial pressure (ICP) may decrease or even halt cerebral perfusion, resulting in secondary ischemic damage to brain cells. Monitoring of ICP, and subsequent goal-directed therapies of ICP and cerebral perfusion, have therefore become a mainstay of neurointensive care. Monitoring of ICP requires neurosurgical placement of pressure transducers within the brain parenchyma, or intraventricular drains. These interventions carry inherent risks, such as bleedings and infections, and they can only be performed in a limited number of hospitals. Noninvasive alternatives to invasive ICP monitoring have been researched for several decades. The optic nerves (ON) are encased within the optic nerve sheaths (ONS). These sheaths are continuations of the meninges covering the brain. The cerebrospinal fluid (CSF) filled subarachnoid space surrounding the brain is connected to a CSF filled ON subarachnoid space (ONSAS). When ICP increases this results in propagation of CSF into the ONSAS. This increases CSF pressure within the ONSAS and leads to dilation of the ONS diameter (ONSD). The ONSD has a wellestablished association with ICP and can be measured with CT, MRI or ultrasound. The ONSD has been used as a noninvasive estimate of ICP for a few decades. Sonographic ONSD measurement is easily performed at the bedside. It is potentially a safe, repeatable, and fast method to screen for elevated ICP. Still, it lacks standardization, is performed with discrepant methodologies, and several important questions regarding ONSD and ICP remain unanswered.

Aims: This thesis aimed to explore some of the less researched issues of ONSD sonography: • What is the inter-rater reliability of ONSD sonography? • Should ONSD be measured external or internal of the dura mater (DM)? • Does correction of ONSD for eye diameter (ED) improve diagnostic accuracy in predicting elevated ICP? • Does diagnostic accuracy and characteristics of ONSD derived predictors of elevated ICP differ between women and men? • Which components of the ONS are affected by ICP?

Methods: In study 1, two raters performed paired, consecutive measurements of ONSD external of the DM (ONSDext), internal of the DM (ONSDint), of the optic nerve diameter (OND) and finally of the ED. Measurements were performed in 20 patients. Inter- and intra-rater reliability was estimated with intra-class correlation (ICC) and compared between ONSDext and ONSDint derived measurements. In study 2-4, the raters independently gathered data on ONSDext, ONSDint, OND and ED in patients with invasive monitoring of ICP. Study 2 compared diagnostic accuracy between ONSDext and ONSDint, with and without correction for ED, by comparing areas under the receiver operator characteristics curve (AUROC). Study 3 reported AUROCs and diagnostic characteristics for these different ONSD derived ICP estimates stratified by sex. In study 4, three distinguishable compartments of the ONS were separated and measured: the OND, the ONSAS thickness and the DM thickness. These were explored for associations with changes in ICP.

Results: Study 1 showed that ONSD can be measured with a good to excellent inter- and intra-rater reliability and a very low risk of inter-rater bias. ICC was significantly better for ONSDext derived measurements than for ONSDint derived measurements. Study 2 showed that correction of ONSD for ED yielded a better diagnostic accuracy. ONSDext derived measurements and ONSDint derived measurements differed significantly but did not yield significantly different diagnostic accuracies. In study 3, elevated ICP could be predicted in women with all explored ONSD derived measurements. In men none of the ONSD derived measurements could predict elevated ICP with an AUROC significantly larger than 0.5. In study 4, ONSAS was the only ONS compartment that independently could predict changes in ICP. Still, both OND and DM added predictive value in multivariable modelling.

Discussion: Study 1 shows that ONSD derived measurements can be performed with an excellent inter-rater reliability, with a standardized protocol. Study 2 shows that ONSDext and ONSDint differ significantly and are not interchangeable. This clearly shows the need for a standardization of ONSD sonography for ICP estimation. With no significant differences in diagnostic accuracy, it could be argued that ONSDext may be preferred based on the significantly better interrater reliability in study 1, though further studies are required before evidencebased recommendations can be stated. Further, study 2, in line with a few previous studies, suggests that correction of ONSD for ED increases diagnostic accuracy and should be used. The results from study 3 complicate matters of ONSD for estimation of ICP. It shows sex differences in diagnostic accuracy, to the degree that ONSD derived measurements may be unreliable predictors of elevated ICP in men. These findings are novel and need to be either corroborated or refuted in larger studies. Results from study 4 complicate the association between ICP and ONSD even further. It challenges current understanding that the ONSAS is the only ONS compartment affected by ICP. It suggests that all compartments of the ONS may be affected by elevated ICP. Potential mechanisms for these phenomena are discussed in the thesis. These new findings warrant further studies as they may lead to a new understanding of the ONSD/ICP mechanism.