Pulmonary hypertension in infants with bronchopulmonary dysplasia: risk factors, mortality and duration of hospitalisation

Introduction

The survival rates of extremely premature infants have improved significantly in the last few decades, but are associated with long term morbidities [1], [2], [3]. Bronchopulmonary dysplasia (BPD) is the most common adverse outcome and can result in complications which may extend into adulthood [4]. One of the significant contributory factors for morbidity and mortality in infants with BPD is the development of pulmonary hypertension (PH) [5].

The incidence of pulmonary hypertension in infants with BPD ranges from 17 to 37% [6], [7], [8]. The pathogenesis of pulmonary vascular disease associated with BPD is multifactorial and can be attributed to the interaction of maternal, genetic and postnatal factors. Inhibition of angiogenesis has been associated with decreased alveolarization and pulmonary vascular growth in an animal model [9]. The arrest of vascular growth and the reduction in the alveolar-capillary surface result in impaired gas exchange. Thus, the lungs are exposed to prolonged ventilation and high oxygen levels along with hemodynamic and inflammatory stressors. This causes pulmonary arterial remodelling leading to fibrosis of the vessel walls [10]. The risk factors associated with the development of pulmonary hypertension include intrauterine growth restriction, preeclampsia, maternal substance abuse, lower gestational age and birthweight, patent ductus arteriosus (PDA) requiring treatment and sepsis [8, 11], [12], [13], [14].

The gold standard to diagnose pulmonary hypertension is cardiac catheterization, but this is often difficult to perform in BPD infants who may have significant vascular and haemodynamic complications [15]. Two-dimensional echocardiography, however, is a valuable tool for screening and diagnosis of PH in such patients [10]. Various studies have investigated the risk factors, outcomes [5], [6], [7], [8, 11, 14, 16], [17], [18], [19], [20], [21] and evolution of pulmonary hypertension in BPD infants [7, 8, 14, 19], [20], [21]. The identified risk factors include early pulmonary hypertension, postnatal steroid administration, tracheostomy and tracheitis [5, 16, 17]. Outcomes associated with PH in BPD infants include increased mortality and morbidity including a prolonged duration of invasive ventilation, increased occurrence of severe IVH and prolonged hospitalization [5, 7, 8, 14, 19], [20], [21]. Our aim was to identify, in infants with BPD, the effect of PH on health-care utilisation and the health related cost of care.

Materials and methods

Results

A total of 428 premature infants were admitted to the neonatal unit during the study period, 211 (49.3%) subsequently developed BPD; 189 infants were included in the analysis and 29 infants were excluded (Figure 1).

Figure 1:

Flowchart of selected cases.

The included infants had a median (interquartile range) gestational age of 25.9 (24.9–27.7) weeks and birthweight of 769 (667–930) grams (Table 2). Twenty-one (11.5%) infants had early PH of which six (28.5%) subsequently developed late PH. Twenty-two (12%) infants had late PH. Amongst the 22 infants with late PH, pulmonary hypertension resolved in 10 infants, four had pulmonary hypertension at discharge from neonatal care and eight died with pulmonary hypertension (Figure 2).

Figure 2:

Development and evolution of pulmonary hypertension during neonatal stay.

The characteristics of infants who developed early PH, late PH, both early and late and no PH have been detailed in Table 3. The infants who developed both early and late PH had a lower (median) gestational age (24.4 weeks) compared to those who had only early PH (25.7 weeks), late PH (24.6 weeks) or no PH (26.1 weeks, p=0.02). Furthermore, infants who developed both early and late PH had a lower (median) birthweight (603 g) compared to those who had early PH (780 g), late PH (673 g) or no PH (818 g, p<0.001).

The infants who developed late PH had a lower gestational age (24.6 vs. 26.1 weeks, p=0.002) and lower birthweight (642.5 vs. 816.5 g, p<0.001) compared to those who did not have late PH. The infants who developed late PH were more often mechanically ventilated on day 28 after birth compared to the ones who did not develop late PH [100 vs. 51.9%, p<0.001; OR: 19.5 (CI: 2.6–148)]. The infants who developed late PH had a longer median duration of mechanical ventilation compared to the infants without late PH (53.5 vs. 27.5 days, p<0.001) [Table 4].

Eleven infants (6%) died in the study population. Eight had pulmonary hypertension at death, of those eight, six had developed PH within 28 days after birth (Figure 2). The presence of late PH was independently associated with increased mortality [OR: 21.4 (4.1–111.3), adjusted p=0.001]. Similarly, the presence of late PH in surviving infants was independently associated with a longer duration of hospital stay [147 vs. 109 days, adjusted p: 0.03] (Table 5).

The total cost of stay for the 182 infants was £17,164,290 ($23,957,659). The median (IQR) total cost of stay in infants with late PH was £113,494 (72,650–169,798) compared to £78,677 (48,476–115,483) in infants with BPD, but no late PH after adjusting for gestational age (p=0.016) [Table 5]. The median cost of stay in infants with BPD and PH was approximately £35,000 higher (44% higher) compared to that of infants with BPD without PH.

The median length of hospitalisation was similar in infants whom PH resolved before discharge [147 (118–189) days] compared to the infants in whom PH did not resolve before discharge or died with PH [147 (72–259) days, p=0.81]. The total cost of stay was not significantly different in infants in whom PH resolved before discharge [£85,992 (67,524–121,939)] compared to the infants in whom PH did not resolve before discharge [£148,458 (80,185–252,455), p=0.07].

Discussion

We have demonstrated late PH was associated with a higher mortality and a longer duration of stay compared to infants with BPD without PH. Furthermore, we found that the total cost of stay for infants with late PH and BPD was higher compared to those who did not develop late PH.

Our study showed that while the incidence of overall mortality was 6% in the study population, the presence of late PH was associated with a 36% increase in mortality. Studies have described the association of PH and mortality with an incidence ranging from 10 to 38% [7, 8, 14, 17, 18, 27]. Notably, among the eight infants who died with pulmonary hypertension, six of them had developed PH within 28 days (Figure 2). Two studies have shown that early PH increases the risk for severe BPD and development of late PH in infants with BPD, but did not describe the impact of this on mortality [16, 17]. This highlights the need for early as well as regular screening for pulmonary hypertension via echocardiographic assessment in infants with BPD.

One study reported the healthcare utilization of infants with PH associated with BPD presenting after two months of age into a BPD clinic and found that the infants were hospitalised for a longer duration and had more respiratory morbidity than infants with BPD but no PH [28]. Our study reported on the total cost of stay based on national average in infants with BPD and PH and to our knowledge this is the first study which specifically looked into the effect of pulmonary hypertension in infants with bronchopulmonary dysplasia on health care costs.

Our study has strengths and some limitations. In our study population we used echocardiography to identify pulmonary hypertension; echocardiography is considered to be an effective and reliable screening tool for detection of pulmonary hypertension in infants with BPD [29]. A limitation, however, was that the infants had echocardiograms at various time intervals according to clinical need rather than at specific intervals, but this makes our results generalisable to clinical practice.

In conclusion, infants with late PH had a longer duration of hospitalization and increased mortality compared to the infants without late PH. The presence of late PH in infants with BPD resulted in increased healthcare costs.