P rimary pontine hemorrhage (PPH) is the most devastating type of intracranial hemorrhage (ICH), with an acute mortality ranging from 30% to 60%. [1][2][3] Various factors including coma at admission, location, and volume of the hematoma were found to associate with these diverse outcomes. However, inconsistent predictors were reported when different parameters and populations were brought into analysis. [4][5][6][7] As a corollary, it would make cogent sense by combining significant factors into a grading scale to enhance predictive power and, meanwhile, provide a useful tool for physicians in decision-making when facing such patients. 8 Currently, there is no standard, widely accepted early prognostic model or clinical grading scale for outcome prediction in PPH patients. The ICH score, composed of age, Glasgow Coma Scale (GCS), infratentorial origin, intraventricular hemorrhage, and hemorrhage volume, has been widely validated in predicting acute mortality, as well as long-term functional outcome in spontaneous ICH.9,10 Easy to use though, the ICH score and its derivatives tended to deem infratentorial hemorrhage as an independent predictor of poor outcome. 11,12 Because of the small structure of the pons but severe manifestations caused by hemorrhagic impairment, the cutoff value Background and Purpose-We aimed to develop and validate a grading scale for predicting 30-day mortality and 90-day functional outcome in patients with primary pontine hemorrhage (PPH). Methods-We retrospectively reviewed records of consecutive patients with first-ever pontine hemorrhage from 3 teaching hospitals between 2005 and 2012. Independent factors associated with 30-day mortality were identified by logistic regression to establish a risk stratification scale, named the new PPH score. For validation of the new PPH score, we prospectively recruited subjects from 10 units between December 2014 and November 2015. The performance of the new PPH score was presented as discrimination and calibration, measured by area under the curve of the receiver operating characteristic and Hosmer-Lemeshow goodness-of-fit, respectively. Results-Data of 171 patients were available for scale development. The new PPH score consisted of 2 independent factors with individual points assigned as follows: Glasgow Coma Scale score 3 to 4 (=2 points), 5 to 7 (=1 point), and 8 to 15 (=0 point); PPH volume >10 mL (=2 points), 5 to 10 mL (=1 point), and <5 mL (=0 point). An independent cohort of 98 patients was applied as an external validation of the new PPH score. Results showed that the new PPH score was discriminative in predicting both 30-day mortality (area under the curve, 0.902) and 90-day good outcome (area under the curve, 0.927). Furthermore, the new PPH score revealed a good calibration (χ 2 =1.387; P=0.846) in 30-day mortality prediction.
Conclusions-The
