Optic nerve sheath diameter and eyeball transverse diameter as a useful tool for the clinical prognosis in patients with stroke during hospitalization / 中华危重病急救医学

ABSTRACT

Objective To evaluate the prognostic value of ultrasonographic measurement of optic nerve sheath diameter (ONSD) and ONSD/eyeball transverse diameter (ETD) ratio in stroke patients during hospitalization. Methods Adult patients with stroke (ischemic stroke or hemorrhagic stroke) admitted to department of critical care medicine of Shanghai General Hospital, Shanghai Jiao Tong University School of Medicine from November 2017 to November 2018 were enrolled. On the day of admission, ONSD and ETD (retina-retina at 3 mm behind the globe along) were detected by ultrasound, the ONSD/ETD ratio was calculated, and the baseline data and outcomes were recorded. Patients were divided into survival group and death group according to their survival status. Locally weighted scatterplot smoothing (LOWESS) and the receiver operating characteristic (ROC) curve were used to calculate the thresholds of ONSD and ONSD/ETD. The correlation between ONSD, ONSD/ETD and prognosis were assessed. Results Thirty-eight of 83 patients (45.8%) survived and were discharged from the hospital, while 45 patients died (54.2%). There were significant differences in Glasgow coma score (GCS), shifting of the middle structure, ONSD and ONSD/ETD between the death group and the survival group [GCS: 4.7±2.8 vs. 11.0±3.2, shifting of the middle structure (mm): 5.8±5.9 vs. 1.3±2.6, ONSD (mm): 5.5±0.4 vs. 4.4±0.5, ONSD/ETD: 0.25±0.02 vs. 0.20±0.02, all P < 0.05]. LOWESS and ROC curve analysis suggested thresholds of ONSD and ONSD/ETD for predicting adverse prognosis of stroke patients were 5.0 mm and 0.25, respectively. By adjusting the influence of confounding factors on prognosis, a prediction model based on ONSD was established, and the ROC curve was drawn. The area under the curve (AUC) was 0.978, the optimal predictive point of the model was 0.870, the sensitivity was 89%, and the specificity was 100%. The ONSD/ETD prediction model was also obtained, and the AUC was 0.988, the optimal prediction threshold of the model was 0.768, and the sensitivity for predicting adverse clinical prognosis was 94%, and the specificity was 97%. The stability of ONSD/ETD was better than that of ONSD. ONSD coefficient of variation was 0.14, and ONSD/ETD coefficient of variation was 0.13. Conclusions ONSD and ONSD/ETD were significantly correlated with the prognosis of critical patients with stroke. The mortality increased significantly in patients with an ONSD greater than 5.0 mm and ONSD/ETD greater than 0.25. ONSD and ONSD/ETD may be promising tools for early assessment of clinical outcomes in these patients.