Objective    To explore the risk factors that affect the mortality of neonates with pulmonary haemorrhage (PH) and establish a mortality prediction score model.     Methods    Totally 244 PH infants were analyzed retrospectively. The risk factors were screened out by logistic regression analysis and the scores were developed according to regression coefficients. Then the receiver operating characteristic curve was constructed and the cutoff was determined.     Results   

The mortality was 74.59% (182/244)

the mean gestational age and median age were (36.513±3.805) weeks and 0.24 days in the survived infants

(35.746±3.929) weeks and 0.33 days in the died infants respectively

without significance between two groups (P0.05). The rate of twins or triplets and birth weight of the survived infants were 3.2% (2/62)

(2 786.866±606.991) g

while 20.1% (25/182)

(2 338.812±756.790) g in the died infants (P0.05). The survived infants had no difference from those died in the rate of neonatal respiratory distress syndrome (43.5% vs 34.6%

χ

2

=1.585

P=0.208)

respiratory failure (77.4% vs 78.6%

χ

2

=0.036

P=0.849)

or cold injury syndrome (35.5% vs 44.5%

χ

2

=0.396

P=0.529). The survived infants had significant difference from those died in the rate of IVH (25.8% vs 53.8%

χ

2

=14.607

P=0.000)

heart failure (22.6% vs 48.9%

χ

2

=13.134

P=0.000)

sepsis (3.2% vs 16.5%

χ

2

=7.134

P=0.008). We further used logistic regression analysis to test above significantly different variables

then the four variables (birth weight

IVH

heart failure and sepsis) were in the logistic regression equation. The predicted percentage of correct was 80.7%. According to regression coefficients

a score ranging from 4 to 40 was developed

the area under the receiver operating characteristic curve was 0.786. At a cutoff point of 9

the pulmonary hemorrhage was predicted with the sensitivity of 0.846

the specificity of 0.661

the positive predictive value of 0.88 and the negative predictive value of 0.594. The mortality was higher in the “low risk group” (the score≤9) than that in the “high risk group” (the score9) (40.6% vs 88%

χ

2

=58.714

P=0.000).      Conclusion    Low birth weight

IVH

heart failure and sepsis are the risk factors in the prognosis of neonates with pulmonary haemorrhage. The infants with the score higher than 9 are assessed to be at “high risk” for death.