Enhancing the prediction for shunt-dependent hydrocephalus after aneurysmal subarachnoid hemorrhage using a machine learning approach.

Early and reliable prediction of shunt-dependent hydrocephalus (SDHC) after aneurysmal subarachnoid hemorrhage (aSAH) may decrease the duration of in-hospital stay and reduce the risk of catheter-associated meningitis. Machine learning (ML) may improve predictions of SDHC in comparison to traditional non-ML methods. ML models were trained for CHESS and SDASH and two combined individual feature sets with clinical, radiographic, and laboratory variables. Seven different algorithms were used including three types of generalized linear models (GLM) as well as a tree boosting (CatBoost) algorithm, a Naive Bayes (NB) classifier, and a multilayer perceptron (MLP) artificial neural net. The discrimination of the area under the curve (AUC) was classified (0.7 ≤ AUC < 0.8, acceptable; 0.8 ≤ AUC < 0.9, excellent; AUC ≥ 0.9, outstanding). Of the 292 patients included with aSAH, 28.8% (n = 84) developed SDHC. Non-ML-based prediction of SDHC produced an acceptable performance with AUC values of 0.77 (CHESS) and 0.78 (SDASH). Using combined feature sets with more complex variables included than those incorporated in the scores, the ML models NB and MLP reached excellent performances, with an AUC of 0.80, respectively. After adding the amount of CSF drained within the first 14 days as a late feature to ML-based prediction, excellent performances were reached in the MLP (AUC 0.81), NB (AUC 0.80), and tree boosting model (AUC 0.81). ML models may enable clinicians to reliably predict the risk of SDHC after aSAH based exclusively on admission data. Future ML models may help optimize the management of SDHC in aSAH by avoiding delays in clinical decision-making.

Aneurysmal subarachnoid hemorrhage in Germany between 2005 and 2018: A nationwide observational study.

BACKGROUND AND PURPOSE: In recent decades, the global incidence of aneurysmal subarachnoid hemorrhage (aSAH) has declined. However, significant regional differences exist. We present the first comprehensive analysis of time trends in the incidence of aSAH and case fatality in Germany. METHODS: All patients hospitalized with aSAH in Germany between 2005 and 2018 were identified using International Classification of Disease, 10th Revision codes I60.0-I60.7. RESULTS: A total of 101,105 cases were included. The incidence of aSAH in Germany decreased at a mean annual rate of 0.5% (p = 0.003) from 8.9 per 100,000 population in 2005 to 8.2 in 2018. Over time, incidences of aSAH declined among patients younger than 55 years and patients aged 70-79 years, increased among ages 60-64 years, and remained stable in age groups 65-69 and 85-89 years. This corresponded to an increase in the mean age of aSAH onset from 55.6 (±14.3) to 59.0 (±14.0) years. Throughout the study period, the mean age of aSAH onset was higher in women compared to men (58.3 ± 14.4 years vs. 56.1 ± 14.1 years). There were no changes in in-hospital case fatality (16.2 vs. 16.6%, p = 0.18), but the duration of hospital stay increased significantly from 19.7 to 24.8 days (p < 0.001). The most frequent aSAH-associated aneurysm location was the anterior circulation throughout the entire study period. CONCLUSIONS: In Germany, the incidence of aSAH decreased between 2005 and 2018, especially in younger parts of the population. This may reflect effects of lifestyle adjustments most pronounced in younger age groups.

A novel score to predict shunt dependency after aneurysmal subarachnoid hemorrhage.

OBJECTIVE Feasible clinical scores for predicting shunt-dependent hydrocephalus (SDHC) after aneurysmal subarachnoid hemorrhage (aSAH) are scarce. The chronic hydrocephalus ensuing from SAH score (CHESS) was introduced in 2015 and has a high predictive value for SDHC. Although this score is easy to calculate, several early clinical and radiological factors are required. The authors designed the retrospective analysis described here for external CHESS validation and determination of predictive values for the radiographic Barrow Neurological Institute (BNI) scoring system and a new simplified combined scoring system. METHODS Consecutive data of 314 patients with aSAH were retrospectively analyzed with respect to CHESS parameters and BNI score. A new score, the shunt dependency in aSAH (SDASH) score, was calculated from independent risk factors identified with multivariate analysis. RESULTS Two hundred twenty-five patients survived the initial phase after the hemorrhage, and 27.1% of these patients developed SDHC. The SDASH score was developed from results of multivariate analysis, which revealed acute hydrocephalus (aHP), a BNI score of ≥ 3, and a Hunt and Hess (HH) grade of ≥ 4 to be independent risk factors for SDHC (ORs 5.709 [aHP], 6.804 [BNI], and 4.122 [HH]; p < 0.001). All 3 SDHC scores tested (CHESS, BNI, and SDASH) reliably predicted chronic hydrocephalus (ORs 1.533 [CHESS], 2.021 [BNI], and 2.496 [SDASH]; p ≤ 0.001). Areas under the receiver operating curve (AUROC) for CHESS and SDASH were comparable (0.769 vs 0.785, respectively; p = 0.447), but the CHESS and SDASH scores were superior to the BNI grading system for predicting SDHC (BNI AUROC 0.649; p = 0.014 and 0.001, respectively). In contrast to CHESS and BNI scores, an increase in the SDASH score coincided with a monotonous increase in the risk of developing SDHC. CONCLUSIONS The newly developed SDASH score is a reliable tool for predicting SDHC. It contains fewer factors and is more intuitive than existing scores that were shown to predict SDHC. A prospective score evaluation is needed.