Article Figures & Data
- Figure 1
Receiver operating characteristic (ROC) analysis for day 90 modified Rankin scale of >3, day 4 death and acute neurological deterioration. Similar area under the curves was observed for ICH volume (ICHV) (semiautomated segmentation and ABC/2), ICH volume/ICV ratio (semiautomated segmentation and ABC/XYZ) and ICH volume/CPV ratio (semiautomated segmentation). AUROC, area under the receiver operating characteristics curve; CPV, cerebral parenchymal volume; ICH, intracerebral haemorrhage; ICV, intracranial volume; SAS, semiautomated segmentation.
- Figure 2
Receiver operating characteristic (ROC) analysis of ambient cistern score and midline shift for day 90 modified Rankin scale of >3, day 4 death and acute neurological deterioration. Ambient Cistern Score had slightly better accuracy than midline shift for prediction of day 4 death and neurological deterioration while midline shift was marginally better in predicting day 90 modified Rankin scale of >3. AUROC, area under the receiver operating characteristics curve.
- Figure 3
Receiver operating characteristic (ROC) analysis of brain atrophy/small vessel disease markers for day 90 modified Rankin scale of >3, day 4 death and acute neurological deterioration. The addition of brain atrophy/small vessel disease markers to known prognostic factors (age, sex, onset to CT time, treatment with GTN, Glasgow Coma Scale, systolic blood pressure and ICH volume) did not improve the accuracy of outcome prediction. AUROC of models including these prognostic factors but without brain atrophy/small vessel disease markers=0.86, 0.85 and 0.85, respectively, for day 90 modified Rankin scale of >3, day 4 death and acute neurological deterioration, which was not very different from those with brain atrophy/small vessel disease markers. AUROC, area under the receiver operating characteristics curve; CPV, cerebral parenchymal volume; GTN, glyceryl trinitrate; ICH, intracerebral haemorrhage; ICV, intracranial volume.
- Figure 4
Scatterplot showing predicted probabilities of haemorrhage expansion with increasing ICHV, ICHV/ICV and ICHV/CPV. The risk of haemorrhage expansion increased with larger ICH volumes but plateaued at 60 mL, when the risk is no longer significant. similar observations were seen with ICH volume/ICV and ICH volume/CPV ratios at cut-offs of >4%. *Models were adjusted for systolic blood pressure, time from stroke onset-to-CT scan and treatment with GTN. CPV, cerebral parenchymal volume; GTN, glyceryl trinitrate; ICHV, intracerebral haemorrhage volume; ICV, intracranial volume.
- Table 1
Baseline characteristics
Baseline characteristics Data Age (years) 73.2 (13.1) Sex, male (%) 73 (55) FAST score=3 (%) 101 (76) Glasgow coma scale <14 (%) 44 (33) National Institutes of Health Stroke Scale (/42) 16 (10, 21) Systolic BP (mm Hg) 175.5 (27.7) Time, from onset to CT scan (hours) 2.3 (1.8, 2.8) Pre-morbid modified Rankin scale >2 (%) 15 (11) Hypertension (%) 75 (57) Diabetes mellitus (%) 20 (15) Ischaemic heart disease (%) 14 (11) Atrial fibrillation (%) 13 (12) Previous stroke (%) 25 (19) Adjudicated CT scan findings Haemorrhage location (%) MCA 103 (78) ACA 4 (3) PCA 7 (5) Lacunar that is, small subcortical stroke 6 (5) Brainstem and cerebellum 15 (11) Midline shift ≥5 mm (%) 24 (18) Leukoaraiosis (%) 68 (51) Cerebral atrophy (%) 123 (94) Old infarcts (%) 87 (65) Intraventricular haemorrhage (%) 54 (41) Haemorrhage longest diameter (cm) <3 44 (33) 3–5 43 (33) 5–8 32 (24) >8 13 (10) Measured CT scan findings ICH volume (SAS, mL) (mean, SD) 37.75 (38.58) IVH volume (SAS, mL) (mean, SD) 8.33 (18.7) Intracranial volume (SAS, mL) (mean, SD) 1419.64 (197.0) Cerebral parenchymal volume (mL) (mean, SD) 1182.0 (197.3) CSF volume (mL) (mean, SD) 240.66 (154.1) Intercaudate distance (mm) (mean, SD) 25.81 (9.7) Sylvian fissure ratio (mean, SD) 0.07 (0.09) van Swieten score (/4)(median, IQR) 2.0 (1.0, 3.0) Ambient cistern score (/4) 0.0 (0.0, 0.0) Data are number (%), mean (SD) or median (IQR).
ACA, anterior cerebral artery; BP, blood pressure; CSF, cerebrospinal fluid; FAST, fast-arm-speech-time; ICH, intracerebral haemorrhage; IVH, intraventricular haemorrhage; MCA, middle cerebral artery; PCA, posterior cerebral artery; SAS, semiautomated segmentation.
- Table 2
The effects of baseline ICH volume (ICHV), ICH volume as a ratio of ICV (ICHV/ICV), ICH volume as a ratio of CPV (ICHV/CPV) (all measured by SAS), midline shift and cistern effacement on day 90 mRS, day 4 death and acute neurological deterioration
Variable Day 90 mRS† Day 4 death Acute neurological deterioration* OR (95% CI) P value OR (95% CI) P value OR (95% CI) P value ICHV (per 10 mL) 1.47 (1.14 to 1.88) 0.003 1.15 (1.01 to 1.29) 0.031 1.20 (1.01 to 1.42) 0.036 ICHV/ICV
(per 1% increase)1.79 (1.24 to 2.59) 0.002 1.24 (1.06 to 1.47) 0.009 1.40 (1.11 to 1.78) 0.005 ICHV/CPV (per 1% increase) 1.63 (1.20 to 2.21) 0.006 1.21 (1.06 to 1.37) 0.005 1.33 (1.09 to 1.62) 0.005 Midline shift (per mm increase) 1.46 (1.11 to 1.93) 0.007 1.17 (1.04 to 1.31) 0.009 1.36 (1.06 to 1.73) 0.014 Cistern effacement (per one point increase from total of 4) 1.76 (1.01 to 3.09) 0.048 2.28 (1.53 to 3.41) <0.001 2.03 (1.10 to 3.76) 0.024 Comparisons by binary logistic regression shown as OR with 95% CIs, with adjustment for age, sex, GCS, systolic blood pressure, time from stroke onset to CT scan, treatment assignment (GTN vs no GTN).
*Acute neurological deterioration which was defined as an increase in National Institutes of Health Stroke Scale of ≥4 or a decrease in GCS of ≥2.25
†OR of mRS score >3
CPV, cerebral parenchymal volume; GCS, Glasgow Coma Scale; GTN, glyceryl trinitrate; ICHV, ICH volume; ICV, intracranial volume; mRS, modified Rankin scale.
- Table 3
The relationships between midline shift, cistern effacement, small vessel disease (old infarcts, van Swieten score) and measures of brain atrophy with MRS at day 90, death at day 4 and acute neurological deterioration
Variable Day 90 mRS† Day 4 death Acute
Neurological deterioration*OR P value OR P value OR Pvalue Small vessel disease Old infarcts (yes vs no) 1.11 (0.39 to 3.16) 0.84 1.80 (0.58 to 5.54) 0.30 0.94 (0.28 to 3.12) 0.91 Leukoaraiosis
(per one point increase in van Swieten score)0.84 (0.58 to 1.23) 0.37 1.23 (0.83 to 1.82) 0.29 1.16 (0.74 to 1.81) 0.52 Brain atrophy Sylvian Fissure Ratio (per 1% increase) 1.06 (0.93 to 1.20) 0.42 0.97 (0.86 to 1.10) 0.68 1.07 (0.93 to 1.23) 0.36 Intercaudate distance (per mm increase) 1.02 (0.96 to 1.08) 0.53 1.03 (0.98 to 1.08) 0.28 1.01 (0.94 to 1.09) 0.72 Cerebral parenchymal volume/Intracranial volume
(per 1% increase)1.01 (0.97 to 1.05) 0.70 1.00 (0.96 to 1.05) 0.93 1.02 (0.97 to 1.08) 0.47 Comparisons by binary logistic regression shown as OR with 95% CIs, with adjustment for age, sex, GCS, systolic blood pressure, time from stroke onset to CT scan, treatment assignment (GTN vs no GTN) and haemorrhage volume.
*Acute neurological deterioration which was defined as an increase in National Institutes of Health Stroke Scale of ≥4 or a decrease in GCS of ≥2.25
†OR of mRS score >3.
CPV, cerebral parenchymal volume; GCS, Glasgow Coma Scale; GTN, glyceryl trinitrate; ICV, intracranial volume; mRS, modified Rankin scale; SFR, Sylvian fissure ratio.
Supplementary Materials
Supplementary data
Supplementary data
Additional Files
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
