The connection between acute severity of the disease value of Oxford and mortality in critical hemorrhagic stroke patients: a retrospective cohort study

Basic features of patients

A total of 1838 hemorrhagic stroke patients were included in the intensive care unit, whereby 1279 survivors and 559 non-peculiarities of the oasis were 33 (27–39). The average age of the patient was 66.57 years (54.08–78.56), 942 patients (52.64%) male. The comorbidities of chronic obstructive lung diseases, coronary heart disease, high blood pressure and diabetes were 0.54%, 10.34%, 58.43%and 19.04%. The 30-day mortality was 30.41% (559 non-survivors and 1279 survivors). The hospital mortality and the mortality rate of the intensive care unit was 25.57% and 20.02%. The length of the stay in the intensive care unit was 3.21 (1.67–8.42) and the length of the hospital stay was 8.67 (4.5-15.58). Further details are shown in Table 1.

Oasis and clinical results hemorrhagic stroke patients

The oasis in the absorption of non-overtaking was significantly higher than that of survivors (39 (34–44) compared to 30 (25–36), P<0.001), the distributions of Oasis and Sapsii with a relevant 30-day mortality were shown in Fig. 1. The 30-day mortality rose with increasing oasis, as well as the 30-day mortality and Sapsii. The age, the initial number of white blood cells, the initial thrombocyte number, the initial hemoglobin concentration, the GCS and the Elixhauser comorbidity index were used in the multivariables regression analyzes according to the univariables logistical regression analysis (shown in Table S5). After the adapted multivariables regression analyzes, the oasis showed a significant connection with the 30-day mortality (or 1.125 per one-point increase, 95%CI [1.107–1.144]Present P<0.0001), intensive mortality (or 1.150 per dot increase, 95% CI [1.128–1.172]Present P<0.001) and hospital mortality (or 1.140 per one-point increase, 95% CI [1.121–1.161]Present P<0.001), further details were displayed in Table 2. The results of the analyzes of SAPS II are shown in Table S6.

Discriminating power of the oasis in hemorrhagic stroke patients

The AUC of the oasis to predict the 30-day mortality was 0.7702 (95% CI [0.748–0.793]), which showed a comparable performance with the Sapsii score (AUC 0.788, 95% CI [0.766–0.810]Present P= 0.096) (Figure S3). The best threshold of the oasis was 35, the specificity and sensitivity was 72.45% or 69.43%, the positive relationship ratio 2.3699 and the relationship of the negative probability was 0.3968 (Table S7). The AUC of the oasis to predict the intensive care unit and hospital mortality was similar to that of Sapsii, as shown in Figure S3. The Kaplan Meier curves showed that a higher Oasis score had predicted a shorter survival time for hemorrhagic stroke patients, as well as the Sapsii (Figure S4).

We further evaluated the performance of the Oasis and Sapsii scales in the prediction of 30-day mortality in intensive care patients (including hemorrhagic stroke and non-hemorrhagic stroke patients in the intensive care unit in a total of 53423 patients).[0.787–0.798] against AUC 0.760, 95%CI[0.754–0.767]Present P= 0.000). This indicates that the forecasting of the Oasis and Sapsii scales differs for the forecast between hemorrhagic stroke patients and intensive care patients (Figure S5).

Sensitive analyzes

We carried out sensitive analyzes in order to evaluate the rod of the results through excluded hemorrhagic stroke patients who are older than 80 years old (> 80 years old). The oasis was still significant with the 30-day mortality (OR = 1.125, 95% CI [1.103–1.146]Present P<0.001), ICU mortality (OR = 1.156, 95% CI [1.131–1.182, p < 0.001), and hospital mortality (OR = 1.144, 95% CI [1.121–1.168, p < 0.001) by the multivariable regression analyses, the results were showed in Table S3. Association of SAPSII scores with outcomes were showed in Tables S8.

Subgroup analyses

The severity of hemorrhagic stroke patients was stratified by GCS, subgroup analysis was performed to evaluate the association of OASIS with 30-day mortality across different patients grouped by GCS. The results were showed in Table S4 and Fig. 2, indicated that the OASIS had similar discriminatory power to the SAPSII for predicting 30-day mortality and ICU mortality expect for predicting 30-day mortality of moderate hemorrhagic stroke patients (GCS 9–12 subgroup), OASIS had lower discriminatory abilities to predict 30-day mortality of GCS 9–12 subgroup than SAPSII (AUC 0.61 with 95% CI [0.53–0.69] against AUC 0.70 with 95% CI [0.63–0.78]Present P= 0.019).

Compare the discriminatory ability of OASIS and SAPSII when recording the forecast of 30-day mortality and the mortality of the intensive care unit layered by GCS. ((A) Comparison of the AUCs for the 30-day mortality by Oasis and Sapsii, which were layered by GCS; ((B) Comparison of AUCs for the ICU mortality by Oasis and Sapsii, which were layered by GCS. Oasis, the acute severity of the disease in Oxford; ICU, intensive care unit; SAPS II, simplified acute physiological score II; AUC, area under the ROC curve; ROC, recipient operating feature.