Predictors of pulmonary hypertension in very low birth weight infants in the neonatal intensive care unit Open Access

Vyas-Read, Shilpa (2016)

Permanent URL:


Approximately 8-23% of premature infants develop pulmonary hypertension (PH), and up to 50% of infants with PH die by 3 years. As a result, professional societies recommend PH screening in premature infants. However, little evidence is available to guide which patients warrant screening, and this void may result in a delay of therapeutic intervention or follow-up.

The two Aims of this proposal were: 1) Determine clinical factors associated with PH 2) Develop a predictive model for PH>30 days of life.

Infants who had the following billing codes: < 32 weeks, birth weight < 1500 grams (BW), neonatal unit, and echocardiograph had records abstracted from a data warehouse. For Aim 1, echocardiographic PH at any point, and for Aim 2, echocardiographic PH >30 days, were the primary outcomes. . For Aim 2, early predictor variables on the outcome of PH> 30 days were evaluated and receiver-operating curves (ROC) were generated to determine cut-points for sensitivity analyses. . Odds ratios and 95% confidence intervals are expressed as (OR, CI) below.

559 infants were included in the overall study, and 92 (16.5%) had PH. For Aim 1, Black race (1.79, 1.02-3.14), atrial septal defect (ASD, 2.72, 1.45-5.11) and patent ductus arteriosus (2.04, 1.19-3.49) increased the odds of PH in multivariable analyses, whereas caffeine decreased the odds (0.49, 0.29 -0.84). 321 patients were included in Aim 2, and a model of birth (birth weight, Apgar 1 minute, Black Race) and early neonatal variables (caffeine, ASD), with positive-pressure ventilation controlled was determined. The ROC showed an area of 0.766, corresponding to a validated sensitivity 80%, specificity 44%, positive predictive value 21%, and a negative predictive value 92%.

Through this proposal, we have identified new factors that contribute to PH, such as race and caffeine. Further, we have developed a model that utilizes early neonatal factors to begin to predict PH. Future steps will include external validation of the model, and implementation of a formalized screening protocol in the neonatal intensive care unit.

Table of Contents

1. Introduction, pages 1-2

2. Background, pages 3-6

3. Methods, pages 7-14

4. Results, pages 15-21

5. Discussion, pages 22-29

6. References, pages 30-33

7. Figure 1, pages 34-35

8. Table 1, pages 36-37

9. Table 2, pages 38-39

10 Table 3, pages 40-41

11. Table 4, pages 42-43

12 Table 5, pages 44-45

13 Table 6, pages 46-48

14 Table 7, page 49

15 Figure 2, page 50

16 Table 8, page 51

17 Table 9, page 52

About this Master's Thesis

Rights statement
  • Permission granted by the author to include this thesis or dissertation in this repository. All rights reserved by the author. Please contact the author for information regarding the reproduction and use of this thesis or dissertation.
  • English
Research Field
Committee Chair / Thesis Advisor
Committee Members
Last modified

Primary PDF

Supplemental Files