Perioperative Variation in Optic Nerve Sheath Diameter – A Prospective Observational Study of Traumatic Brain Injury Patients Undergoing Decompressive Craniectomy
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.355178
Source of Support: None, Conflict of Interest: None
Keywords: Decompressive craniectomy, intracranial pressure, optic nerve sheath diameter, traumatic brain injury, ultrasonography
Rapid industrialization, urbanization, and motorization have added to the momentum of global economic growth. Emerging non-communicable diseases and accidents are 'silent epidemic' to any community undergoing a rapid economic and social transition. Traumatic brain injuries (TBI) are associated with high mortality, morbidity, and socio-economic impact. The estimated all-cause incidence of TBI is 69 million (95% CI 64–74 million) individuals each year. Central Asia, Eastern Europe and Central Europe regions experience the greatest overall burden of TBI. Over the last three decades age-standardized incidence, prevalence, and years of life lived with disability rates due to TBI has significantly increased by 3·6% (95% UI 1·8 to 5·5); 8·4% (95% UI 7·7 to 9·2); 8·5% (7·6–9·3) respectively per 1,00,000 population. The deleterious effects of TBI are almost always attributed to raised intracranial pressure (ICP) and brain parenchymal injury. Measurement of ICP gains relevance in this context.
Various invasive and non-invasive methods are validated to measure ICP. Invasive methods are the gold standard, but associated with procedure-related severe complications like infections, intra-ventricular bleeds and cerebrospinal fluid (CSF) leakage. Measurement of the optic nerve sheath diameter (ONSD) by transbulbar ultrasonography (TBUS) can suffice non-invasive ICP measurement with considerable accuracy.,
Craniectomy is among the third tier and the highest intervention in treatment of raised ICP in severe TBI. Neurosurgical procedures are associated with fluctuations in ICP due to various reasons like CSF loss through dural incisions, use of external ventricular/lumbar drains and/or patient surgical position. These changes can also affect the CSF flow dynamics over the optic nerve sheath (ONS). This can be non-invasively measured by TBUS. Till date no published studies have analyzed the pre/post-operative effect of decompressive craniectomy on CSF flow dynamics measured by TBUS. The primary objective of this study was to evaluate the perioperative variation in ONSD measured by TBUS in TBI patients undergoing emergency decompressive craniectomy for TBI. The secondary objectives were to find out whether there is an intraoperative variation in ONSD based on laterality of injury/side of craniectomy surgery.
After obtaining Institutional Ethics Committee approval (HEC.No. 12/39/2019/MCT dated 20.12.2019 of Human Ethics Committee, Government Medical College, Thiruvananthapuram), we conducted a prospective observational study to evaluate the postoperative variation in ONSD compared to preoperative values, measured by TBUS, in patients undergoing decompressive craniectomy for TBI. The study was conducted from January 2020 to May 2020. Informed written consent was obtained from legal representative of study participants. The study was prospectively registered with the Clinical Trial Registry of India (CTRI/2020/01/022860 dated 20.01.2020). The study was conducted in the emergency neurosurgical operating rooms (OR) of our institution.
A total of 45 consecutive adult patients aged between 18 to 60 years undergoing emergency craniectomy surgery for TBI were recruited to the study. Patients with intraorbital injury, and/or preoperative disorders such as neoplasms, inflammatory diseases, sarcoidosis, pseudo-tumor cerebri, Grave's disease, and extrinsic compression of the optic nerve were excluded from the study. Patients undergoing craniectomy for stroke (ischemic/hemorrhagic) were excluded from recruitment.
Once in the OR, patients were monitored with electrocardiogram, pulse oximetry, non-invasive arterial pressure, and the bispectral index (BIS) monitor. General anesthesia (GA) was induced with intravenous (IV) doses of propofol (2 mg/kg), fentanyl 1–3 mcg/kg followed by administration of vecuronium bromide (0.1 mg/kg). After endotracheal intubation, mechanical ventilation was initiated using anesthesia work-station with 50% oxygen-air mixture, using a tidal volume of 8–10 ml/kg adjusted to a peak inspiratory pressure below 25 cm H2O. Respiratory rate was adjusted to maintain the end-tidal carbon dioxide (etCO2) levels between 30 and 40 mmHg during surgery. Continuous arterial blood pressures were monitored with invasive vascular access in the non-dependant radial artery.
Subsequently, patients underwent bilateral TBUS measurements for ONSD. A layer of sterile water-soluble jelly was applied over the closed upper eye lid of the patient and sonography performed with portable USG machine using high frequency (7-11MHz) linear probe; in transverse-plane [Figure 1] as per standard guidelines. Sterile precautions were maintained during sonography and pressure on eye ball was avoided. The probe was angled medially and caudally to obtain optimal axial view of the optic nerve entering the eye ball, after appropriately adjusting depth and gain of the image. The images were saved from recording of each eye subsequent to which the patients underwent craniectomy surgery. After the conclusion of surgery and before the reversal of neuromuscular blockade second bilateral TBUS was done using the same preoperative protocol.
Modality of maintenance anesthesia – total intravenous anesthesia (TIVA) or inhalational anesthesia or combined; adjusted to maintain a BIS within 40 to 60; was at the discretion of the anesthesiologist conducting the procedure. Also the choice of IV fluids and cerebral decongestants were with the anesthesiologist conducting the procedure.
All TBUS procedures (pre/post-operative) were conducted, identified using code and saved in the sonography machine by the principal investigator throughout the study. The ONSD measurements were done by another investigator blinded to the pre/post-operative nature of the saved image. The ONSD measurements were handed-over written to the principal investigator using the image identification code. Each ONSD measurement was done 3mm behind the eye-ball perpendicular to the line of transection to the eye-ball using electronic calliper[Figure 1]. The average of three recordings each pre/post-operative was analysed for each eye, and average of right and left eye was also calculated. Demographic parameters, intraoperative hemodynamic parameters, choice of anesthesia, cerebral decongestants used, total IV fluids/blood transfusion received and total duration of surgery/anesthesia/ONSD imaging were recorded.
No previous studies have evaluated the before and after effect of craniectomy on ONSD measured by TBUS. The effect size (0.5 mm) was calculated from a previous study measuring ONSD variation before and after ventriculo-peritoneal shunt surgery. The sample size was calculated at two-tailed significance level with alpha error of 0.05, beta error of 20, and power of 80%. The population to be recruited was 40 patients. We recruited 45 cases so as to compensate for outliers and confounders.
Data collected in a prescribed proforma was entered in to a Microsoft Excel spreadsheet and analyzed using SPSS version. 25 (IBM, USA). Bee-swarm plots were constructed using Graph-pad prism software version 8.0.1 (San Diego, CA92018). Results on continuous measurements are presented as mean ± SD; and categorical variables as percentages. Student-t test was used to find the significance of pre and post-operative ONSD measurements. A P value of < 0.05 was considered significant. Correlations between ONSD's of either eye were derived using Pearson correlation coefficient. Bland-Altman plots were used to check the level of agreement between right and left ONSD, both pre/postoperatively.
This prospective observational study was conducted among 45 consecutive adult patients undergoing emergency craniectomy for TBI. Flow diagram showing enrolment, exclusion and analysis of participants through each stage of our study is shown in [Figure 2]. The patient and surgical case characteristics of the study population are as described in [Table 1]. With 4 ONSD measurements examined in each patient a total of 180 TBUS examinations were performed.
Mean age of the study population was 37.76 ± 12.93 years. Eighteen patients underwent right-sided large fronto-temporo-parietal (FTP) craniectomy and 22 cases underwent similar left-sided craniectomy; whereas 5 cases underwent bifrontal craniectomy for non-diffuse primary brain injury (focal hematoma). Based on preoperative Glasgow Coma Scores (GCS), 34 cases (75.5%) had severe TBI (GCS 3-8); 10 cases (22.2%) moderate TBI (GCS 8-12); and 1 case (2.2%) mild TBI (GCS 13-15). Preoperative Marshall CT grade was grade 4; 3; 2; and 6 in 22 (48.8%); 9 (20%); 7 (15.5%) and 7 (15.5%) of the cases respectively. The cases were operated within a time frame of 11.98 ± 4.28 hours after the onset of injury with 34 (75.5%) cases on preoperative mechanical ventilation. Mannitol IV (dose 0.25-1g/kg) was used for osmotherapy in all the cases. All cases were operated in supine position with a mean operative duration of 192.18 ± 27.05 minutes and anesthesia time of 215.84 ± 29.70 minutes. The perioperative TBUS examinations were completed within 11.20 ± 0.93 minutes.
[Figure 3]a, [Figure 3]b, [Figure 3]c, [Figure 3]d shows the evolution of physiological and ventilation variables over the course of the surgery. Mean Cerebral Perfusion Pressure (CPP) derived using the ICP measurement obtained from pre/postoperative ONSD, as in the method described by Robba C et al., (non-invasive ICP = 5 × ONSD-14 [CPP = Mean blood pressure-ICP) is portrayed along with physiologic variables in [Figure 3]a.
Comparison of participant ONSD before and after surgery, and a pooled comparison of ONSD variations in patients undergoing right versus left FTP craniectomy and bifrontal craniectomy are as described in [Table 2]. Average ONSD reduced significantly by 0.249 ± 0.148 mm (P < 0.001) after craniectomy[Figure 4]a and [Figure 4]b. On pooled analysis of cases undergoing right versus left-sided craniectomy average ONSD reduced significantly by 0.252 ± 0.173 mm (P < 0.001) and 0.259 ± 0.139 mm (P < 0.001) respectively [Table 2]. The ONSD reduced by 0.193 ± 0.092 mm (P = 0.009) in cases undergoing bifrontal craniectomy.
On linear regression analysis, ONSD of right eye with left eye and vice-versa were strongly correlated both pre/postoperatively with Pearson correlation coefficients (r)=0.879 (P < 0.001) and r = 0.827 (P < 0.001), respectively [Figure 5]a and [Figure 5]b. Bland-Altman analysis of preoperative right versus left ONSD yielded a mean difference of 0.015 with limits of agreement as -0.401 to 0.432; and 4.4% (2/45) of the plots were outside the limits of agreement [Figure 6]a. Bland-Altman analysis of postoperative right versus left ONSD yielded a mean difference of 0.016. The limits of agreement were -0.532 and 0.500; with 1 (2.2%) plot outside the limits of agreement [Figure 6]b.
This study provides evidence that TBUS measured ONSD is bilaterally increased in patients with TBI undergoing craniectomy irrespective of the laterality of lesions in preoperative CT, and the subsequent side of surgery. There was a significant reduction in ONSD immediately after the surgery; nevertheless, the diameters did not near normal range irrespective of the side of surgery. To our knowledge, till date this is the first study to evaluate a pre-post comparison of TBUS measured ONSD in patients undergoing craniectomy, with pooled analysis based on laterality of surgery.
The CSF within ONS is in direct communication with subarachnoid space and chiasmatic cisterns. TBUS measured ONSD is a validated non-invasive tool to quantify ICP after TBI., There is a 2.0-22.7 fold increased risk of mortality in TBI patients with enlarged ONSD., Similarly enlarged ONSD can predict neurologic outcome on return of spontaneous circulation after cardiac arrest. An ONSD of 5.8mm is often described as appropriate cut-off in TBI. The point-of-care diagnostic benefit and feasibility of rapid interventions with TBUS is unmatched when compared to CT/MRI with comparable ONSD measurements.
The advantage of CSF drainage and invasive ICP monitoring with ventricular catheters cannot be ignored in severe TBI cases with GCS≤6; however not all patients of our study population were suitable candidates for invasive ICP monitoring. Hence, ONSD quantified ICP is salutary in such circumstances. We excluded patients with ischemic/hemorrhagic stroke undergoing craniectomy as multiple co-morbid illnesses in those predominant elderly populations can confound the study methodology. Only transverse TBUS measurements were done and sagittal TBUS avoided in our study considering the emergency nature of the surgery. Axial measurement of ONSD with the TBUS probe transverse to eyelid is most likely to be predictive of high first-measured ICP. Interobserver variability was negated with the same investigator performing TBUS in all the study cases and another blinded investigator performed ONSD measurements of the electronically saved image. Previous studies have distinguished TBUS as a technique easy to learn with low intra/inter-observer variability.,
Our study suggests that neither the CSF loss associated with brain parenchymal dissection in diffuse injury nor the intracranial hypertension lowering effect of craniectomy did contribute to a normalizing ONSD. We postulate four possible reasons attributing to our findings. Firstly TBI represents a state of neurologic hyper-inflammation along with raised ICP and cerebral edema. The linear increase of ONSD in TBI cases should also be attributed to this inflammatory response. ONSD being a papilledema corroborate the toxic milieu due to astrocyte damage in TBI can persist beyond the duration of craniectomy. Secondly, the CSF flow dynamics in small volume ONS (volume of 0.1-0.2 mm3) may not be linearly extrapolated to similar dynamics in ventricles and cisterns in the context of an injured brain. Dural lymphatic clefts in ONS can also alter CSF flow. Thirdly, a heterogeneous higher distribution of lipocalin-like prostaglandin-D synthase (L-PGDS) around the ONS can also contribute to persistent raised ONSD. Fourthly, raised ICP in TBI can compromise pial-septal blood supply to the ONS causing hypoxia and metabolic toxicity that further trigger a vicious cycle of neuro-inflammation. Moreover, the vasoconstrictive effect of L-PGDS and cerebral auto-regulatory failure in TBI can worsen this hypoxia. All these factors could have together contributed to a partial compartmentalization of CSF space within ONS. Hence, the shrinkage of once enlarged ONSD after important increase in ICP in TBI is likely related to multiple factors and not just limited to mere displacement of CSF. In-vitro experiments by Hansen et al. showed that residual dilation of ONSD persisted on decompression from high ICP levels. The modest reductions in ONSD among our patients who underwent bifrontal craniectomy compared to lateral craniectomy could also be attributed to hyperinflammation and partial ICP compartmentalization, with the ONS being more proximal to the site injury in this cohort.
The research till date provides variable evidence on a simultaneous change of TBUS measured ONSD with ICP and CSF drainage. CSF drainage with ventriculo-peritoneal shunt has shown to reduce TBUS measured ONSD immediately after 30 minutes in the study by Choi SH et al., and after 12 hours in the study by Bhandari D et al. Schott CK et al. did not detect any acute change in ONSD among ambulatory patients after elective lumbar puncture. Whereas, Wang L et al. found a significant change in ONSD that correlated with ICP (r = 0.702, P < 0.001) among lumbar puncture cases at 1 month follow-up. Gao Y et al. found no significant correlation (r = 0.205, P = 0.362) between ICP and ONSD performed at the 6th postoperative hour in their small cohort of 33 non-diffuse primary brain injury cases who underwent hemicraniectomy; however, a preoperative TBUS and pooled analysis based on laterality of surgery was not performed. Similarly, ONSD did not normalize after ICP reduction in sub-arachnoid hemorrhage patients.
Outside the scenarios that involve direct drainage of CSF, a significant increase in ONSD has been found during CO2 pneumoperitoneum for laparoscopic pelvic/abdominal/prostate surgery and during tracheal suction. Dynamic alterations of PEEP and etCO2 in anesthetized patients can also alter ONSD. Trendelenberg position has been found to be equivocal in its effects on ONSD.,
Hence, as TBUS measured ONSD gains momentum in multiple clinical scenarios, a delayed return of ONSD to normal range in post-craniectomy patients can debate its diagnostic utility in this selected population. The underlying mechanistic underpinnings that determine the elastic properties of ONS and its subsequent return to normal range can be variable in cases with normal brain and an injured brain.
That we did not perform serial TBUS measurements and correlated it with outcome across mild, moderate and severe TBI is a limitation of our study. Correlation between invasive ICP and ONSD as well could have added further to the lucidity of our results, nevertheless not all cases in our study population qualified for invasive ICP monitoring. Studies in larger samples of TBI cases can validate our results further.
We showed that TBUS measured ONSD is bilaterally increased preoperatively in TBI cases undergoing decompressive craniectomy. TBUS measured ONSD reduces significantly immediately after craniectomy; however, the diameters did not near the normal range. There hold a strong correlation between right/left side ONSD measurements in TBI irrespective of the laterality of injury or side of surgery.Variable elastic properties of the ONS in an injured brain can possibly explain our findings. Post-craniectomy ONSD measurements have many limitations in being truly representative of ICP, hence its value in comparison with ICP monitoring for prognostication need be addressed with caution.
Declaration of patient consent
The authors certify that they have obtained all appropriate patient consent forms. In the form the patient(s) has/have given his/her/their consent for his/her/their images and other clinical information to be reported in the journal. The patients understand that their names and initials will not be published and due efforts will be made to conceal their identity, but anonymity cannot be guaranteed.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6]
[Table 1], [Table 2]