Abstract
Objective
To measure plasma levels of vascular endothelial growth factor (VEGF) and several cytokines (Interleukin [IL]-6 IL-8, IL-10) during the first week of life to examine the relationship between protein expression and likelihood of developing respiratory distress syndrome (RDS) and bronchopulmonary dysplasia (BPD).
Study design
Levels of IL-6, IL-8, IL-10, and VEGF were measured from plasma obtained from preterm patients during the first week of life. Newborns were recruited from a single center between April 2009 and April 2019. Criteria for the study included being inborn, birth weight of less than 1500 grams, and a gestational age of less than 32 weeks at birth.
Results
The development of RDS in preterm newborns was associated with lower levels of VEGF during the first week of life. Higher plasma levels of IL-6 and IL-8 plasma were associated with an increased likelihood and increased severity of BPD at 36 weeks postmenstrual age. In contrast, plasma levels of VEGF, IL-6, IL-8, and IL-10 obtained during the first week of life were not associated with respiratory symptoms and acute care use in young children with BPD in the outpatient setting.
Conclusions
During the first week of life, lower plasma levels of VEGF was associated with the diagnosis of RDS in preterm infants. Preterm infants with higher levels of IL-6 and IL-8 during the first week of life were also more likely to be diagnosed with BPD. These biomarkers may help to predict respiratory morbidities in preterm newborns during their initial hospitalization.
One of the more common complications of preterm birth is respiratory disease with up to 50 000 infants born annually at risk for the archetype of preterm respiratory disease, bronchopulmonary dysplasia (BPD).1 Longitudinal studies in preterm and full-term infants and young children indicate that early life events can have adverse effects on lung function into adulthood.1–3 The ability to predict preterm respiratory outcomes may allow for earlier diagnosis, improved targeting of therapies, and better counselling of families.4 However, it remains difficult to predict which preterm infants in the neonatal intensive care unit (NICU) are more likely to have long-term respiratory disease or symptoms, even among patients with similar clinical risk factors such as gestational age and birth weight.
In this study, we examined the relationship between serum biomarkers and respiratory outcomes in preterm newborns. In particular, vascular endothelial growth factor (VEGF) is important for normal alveolar development. In vitro studies using endothelial progenitor cells from human preterm cord blood revealed that hyperoxia can suppress VEGF signaling and impair endothelial progenitor cell growth.5 Preclinical studies using mice and rats have shown that blockade of VEGF-VEGFR2 signaling can impair endothelial and alveolar growth in the neonatal lung.6,7
Lung inflammation can significantly impair postnatal lung growth; thus, markers of inflammation are an attractive target given the role of inflammation in the development of BPD.8,9 Prior studies of preterm infants have identified associations between the development of BPD at 36 weeks corrected age and serum cytokine levels obtained within the first few weeks of life.10–13 Similarly, associations have been observed between preterm pulmonary vascular disease at 36 weeks corrected age (ie, pulmonary hypertension) and higher plasma levels of Interleukin (IL)-6, IL-8, IL-10, and tumor necrosis factor-α.14
The most commonly used phenotype for preterm respiratory disease is BPD, which is typically defined at 36 weeks corrected age.15,16 However, there are no common definitions for respiratory disease after this time-point; hence, there is a paucity of studies examining the predictive power of early life biomarkers on longer term respiratory outcomes in the preterm population.
Based on prior preclinical studies, we hypothesized that VEGF levels and inflammatory biomarkers in the early neonatal period would be associated with respiratory outcomes in preterm infants. The timepoints for respiratory outcomes included the first week of life (respiratory distress syndrome [RDS]), 36 weeks corrected gestational age (BPD), and infants and toddlers after initial hospital discharge (acute care use and chronic respiratory symptoms before 3 years of age). For this study, we selected IL-6 and IL-8, inflammatory cytokines previously associated with BPD or pulmonary hypertension in prior studies, and IL-10, a protein involved in immunoregulation and inflammation.10–14 We also examined the plasma levels of VEGF, a gene previously associated with the development of BPD and pulmonary hypertension.6,17 To assess for these associations, we used data from preterm infants from an existing biomarker database and also identified infantss who were enrolled subsequently in a separate outpatient registry of preterm infants with BPD, which captures respiratory outcomes outside the NICU.
Methods
For this single-center retrospective cohort study, 447 patients were recruited from the inpatient maternity ward at Johns Hopkins Hospital (Baltimore, MD) between April 2009 and April 2019. Inclusion criteria included being inborn, a birth weight of less than 1500 grams, and a gestational age of less than 32 weeks at birth. This study was approved by the Johns Hopkins University Institutional Review Board (Protocol #: NA_00026068) and written informed consent was obtained from a parent after 2017, the time at which exempt status was changed. A subset of infants was subsequently followed in the Johns Hopkins outpatient pulmonary clinic after discharge with ongoing data collection through questionnaires entered into a separate registry. The outpatient registry was approved by the Johns Hopkins University Institutional Review Board (Protocol #: NA_00051884) and oral informed consent was obtained from parents or guardians.
Biomarker Collection
Specimens were discarded blood obtained for routine laboratory studies for clinical care during the first week of life, including cord blood specimens. For the outcome of RDS only samples from day of life 0 to 1 were used for analysis. Plasma samples were stored at 4°C for no more than 24 hours before aliquoting, then stored at −80°C until assayed at the Allen Everett’s laboratory (Johns Hopkins University). Biomarker assays (for IL-6, IL-8, IL-10, and VEGF) were performed using a custom multiplex ELISA using robotically spotted capture antibodies (Meso Scale Discovery). Full details on the assay have been published previously.18 The lower limits of detection (LLOD) for the IL-6, IL-8, IL-10, and VEGF assays were 0.07 pg/mL, 0.27 pg/mL, 0.06 pg/mL, and 0.16 pg/mL, and the upper limits of detection (ULOD) were 758 pg/mL, 621 pg/mL, 341 pg/mL, and 1340 pg/mL, respectively.
Clinical Data
Demographic and inpatient clinical data for the study population was obtained via chart review. The outcome of RDS was defined as noted by the discharging physician. The outcome and severity of BPD (mild, moderate, severe) was defined by standardized National Heart, Lung and Blood Institute criteria.15 Race/ethnicity were self-reported; for the purposes of analysis, any patients that were reported to have any non-White ancestry were coded as non-White. Outpatient clinical data (acute care use [ie, emergency department visits and hospitalizations for respiratory symptoms, steroid courses, antibiotic courses for respiratory illnesses over the preceding 2 months] and chronic symptoms [ie, cough, congestion, wheezing, coughing with feeds, nighttime cough, use of short-acting beta-agonists at home, tolerance of physical activity over the past week]) were collected via caregiver questionnaire as a convenience sample at routine outpatient visits in a subspecialty BPD clinic. Only questionnaire data obtained before 3 years of age were used.
Statistical Analyses
Biomarkers values below the LLOD and more than twice the ULOD were censored to LLOD and twice the ULOD, respectively (biasing toward the null). Biomarkers were log-transformed owing to skewed distribution and averaged for each patient if the patient had more than one measurement of a specific biomarker (0–1 days for RDS outcomes and 0–7 days for BPD and outpatient outcomes). The χ2 tests, t tests, and ANOVA were used to compare frequencies and means between groups to identify potential confounders. Multivariate logistic regression was used to assess for associations between mean of log biomarker levels (independent variable) and RDS (dependent dichotomous variable) adjusted for confounders. Multivariate ordered logistic regression was used to assess for associations between mean of log biomarker levels (independent variable) and BPD (dependent ordinal variable) adjusted for confounders. Multivariate logistic regression was used to assess for associations between the mean of the log biomarker levels (independent variable) and outpatient respiratory outcomes (dependent dichotomous variables) separately adjusted for confounders and clustered by patient to account for the possibility of having multiple questionnaires (obtained at different timepoints) per patient. For statistical analyses, a P value of less than .05 was considered significant. Stata IC 15 was used for all analyses (StataCorp LLC).
Results
A total of 447 preterm infants were recruited for this study (Table I). Approximately one-half of the patients (48.3%) were female and more than two-thirds (68.0%) were non-White. The mean gestational age was 27.6 ± 2.3 weeks with mean birth weight of 962 ± 272 g. A fraction of patients were reported to have oligohydramnios (9.4%) or intrauterine growth restriction (17.9%).
Table I.
Study population
| Characteristics | Entire population (n = 447) |
|---|---|
| Prenatal | |
| Maternal age (years) | 28.9 ± 6.5 [15–51] |
| Oligohydramnios (yes) | 9.4 |
| Intrauterine growth restriction (yes) | 17.9 |
| Prenatal steroids (yes) | 92.4 |
| Perinatal | |
| Cesarean delivery (yes) | 62.9 |
| Sex (female) | 48.3 |
| Race/ethnicity (non-White) | 68.0 |
| Gestational age (weeks) | 27.6 ± 2.3 [23.1–31.9] |
| Birth weight (g) | 962 ± 272 [420–1490] |
| RDS (yes) | 91.5 |
| BPD Severity (n = 411) | |
| No BPD | 24.1 |
| Mild BPD | 20.2 |
| Moderate BPD | 31.9 |
| Severe BPD | 23.8 |
Values are mean ± SD [range] or percent.
RDS
A majority of patients were diagnosed with RDS (91.5%), and most patients had also received prenatal steroids (92.4%). VEGF levels were lower in those with RDS compared with those without (P < .016) and IL-10 levels tended to be higher in those with RDS (P = .056) (Table II). Untransformed levels without statistical comparisons can be found in Table III (available at www.jpeds.com). There was no difference in IL-6 or IL-8 levels between those with RDS and those without. We compared demographic and clinical factors between those with RDS and those without to identify potential confounders and found that earlier gestational ages and lower birthweights (and potentially male sex) were associated with RDS (Table IV; available at www.jpeds.com). Multivariate logistic regression was used to determine the potential risk of RDS associated with biomarker levels (Table V). After adjusting for sex, gestational age, and birthweight, only VEGF levels were associated with RDS (P = .016). Specifically, because data were log transformed, then each 10-fold increase in mean VEGF level was associated with a 35% decrease in the risk of RDS.
Table II.
Biomarkers and inpatient respiratory outcomes
| Mean of log biomarkers | RDS | BPD | ||||||
|---|---|---|---|---|---|---|---|---|
| Absent (n = 38) | Present (n = 403) | P value | None (n = 99) | Mild (n = 83) | Moderate (n = 131) | Severe (n = 98) | P value | |
| VEGF | 1.70 ± 0.97 [−0.79 to 2.97] | 1.23 ± 1.17 [−0.79 to 3.16] | .016 | 1.79 ± 0.93 [−0.79 to 3.13] | 1.74 ± 1.03 [−0.79 to 3.26] | 1.68 ± 0.86 [−0.79 to 3.21] | 1.42 ± 1.03 [−0.79 to 3.00] | .036 |
| IL-6 | 1.11 ± 0.91 [−0.72 to 3.18] | 1.20 ± 0.88 [−1.14 to 3.18] | .56 | 0.73 ± 0.68 [−1.14 to 3.18] | 0.91 ± 0.63 [−1.14 to 2.57] | 1.04 ± 0.58 [−1.14 to 2.81] | 1.28 ± 0.69 [−1.14 to 3.18] | <.001 |
| IL-8 | 2.08 ± 0.48 [1.07 to 3.09] | 2.06 ± 0.53 [−0.57 to 3.09] | .79 | 1.88 ± 0.43 [−0.57 to 3.09] | 2.02 ± 0.48 [−0.57 to 3.08] | 2.09 ± 0.34 [0.45 to 2.81] | 2.18 ± 0.46 [−0.57 to 3.09] | <.001 |
| IL-10 | 0.22 ± 0.84 [−1.22 to 2.07] | 0.46 ± 0.73 [−1.22 to 2.37] | .056 | −0.12 ± 0.64 [−1.22 to 2.22] | 0.08 ± 0.72 [−1.22 to 1.88] | 0.21 ± 0.66 [−1.22 to 2.17] | 0.42 ± 0.63 [−1.22 to 2.24] | <.001 |
Values are mean ± SD [range]
Table III.
Untransformed biomarkers and inpatient respiratory outcomes
| RDS | BPD | |||||
|---|---|---|---|---|---|---|
| Means of biomarkers | Absent (n = 38) | Present (n = 403) | None (n = 99) | Mild (n = 83) | Moderate (n = 131) | Severe (n = 98) |
| VEGF | 197 ± 259 | 132 ± 236 | 203 ± 253 | 222 ± 317 | 177 ± 281 | 122 ±183 |
| IL-6 | 135 ± 388 | 110 ± 284 | 32 ± 159 | 27 ± 64 | 31 ± 78 | 89 ± 249 |
| IL-8 | 224 ± 293 | 204 ± 249 | 117 ± 176 | 173 ± 224 | 161 ± 125 | 223 ± 199 |
| IL-10 | 8.0 ± 20.1 | 10.3 ± 25.0 | 3.8 ± 17.6 | 5.8 ± 14.6 | 5.2 ± 14.8 | 7.5 ± 19.1 |
Values are mean ± SD.
Table IV.
Study demographics by RDS
| Characteristics | Entire population (n = 447) | No RDS (n = 38) | RDS (n = 409) | P value |
|---|---|---|---|---|
| Maternal age (years) | 28.9 ± 6.5 [15–51] | 29.4 ± 6.3 [16–40] | 28.8 ± 6.6 [15–51] | .59 |
| Oligohydramnios (yes) | 9.4 | 13.2 | 9.1 | .41 |
| Intrauterine growth restriction (yes) | 17.9 | 23.7 | 17.4 | .33 |
| Prenatal steroids (yes) | 92.4 | 97.4 | 91.9 | .23 |
| Cesarean delivery (yes) | 62.9 | 60.5 | 63.1 | .76 |
| Sex (female) | 48.3 | 63.2 | 46.9 | .06 |
| Race/ethnicity (non-White) | 68.0 | 76.3 | 67.2 | .25 |
| Gestational age (weeks) | 27.6 ± 2.3 [23.1–31.9] | 28.9 ± 2.2 [24.3–31.9] | 27.4 ± 2.2 [23.1–31.9] | <.001 |
| Birthweight (g) | 962 ± 272 [420–1490] | 1087 ± 263 [500–1490] | 951 ± 271 [420–1480] | .003 |
Values are mean ± SD [range] or percent.
Table V.
Predicted risk of inpatient respiratory outcomes
| Characteristics | Risk of RDS* (n = 447) | P value | Risk of BPD† (n = 411) | P value |
|---|---|---|---|---|
| VEGF | 0.65 ± 0.12 [0.46–0.92] | .016 | 0.98 ± 0.11 [0.79–1.21] | .84 |
| IL-6 | 0.84 ± 0.18 [0.56–1.28] | .42 | 1.71 ± 0.27 [1.25–2.33] | .001 |
| IL-8 | 0.59 ± 0.22 [0.28–1.24] | .16 | 1.77 ± 0.44 [1.08–2.89] | .024 |
| IL-10 | 1.13 ± 0.29 [0.69–1.86] | .62 | 1.16 ± 0.17 [0.86–1.55] | .33 |
Values are OR ± SE [95% CI].
Logistic regression with independent variable of biomarker (mean log-transformed levels from days 0 to 1 of life) and dependent dichotomous variable of RDS adjusted for sex, gestational age, and birthweight.
Ordered logistic regression with independent variable of biomarker (mean log-transformed levels from days 0 to 7 of life) and dependent categorical variable of BPD severity (0 = no BPD; 1 = mild BPD; 2 = moderate BPD; 3 = BPD) adjusted for gestational age, intrauterine growth restriction, birthweight, RDS, and prenatal steroids. For a 10-fold increase in biomarker level, the odds of severe BPD vs all other categories of BPD is the aOR; likewise, the same OR reflects the risk of the combined mild, moderate, and severe BPD (any BPD) vs no BPD.
BPD
Of the study population of 447 patients, the severity of BPD could only be ascertained for 411. Available medical records did not permit for determining BPD severity for 8 patients, and an additional 28 were deceased before the standardized timepoint for the assessment of BPD; these patients were excluded from BPD analyses. Of these 411 patients, 24.1% did not meet the criteria for BPD, 20.2% had mild BPD, 31.9% had moderate BPD, and 23.8% had severe BPD. Increases in BPD severity were associated with lower levels of VEGF (P = .036) and higher levels of IL-6, IL-8, and IL-10 (all P < .001) (Table II). In terms of potential confounders, we also found that earlier gestational ages, lower birth weights, nonreceipt of prenatal steroids, and prior RDS were associated with increased severity of BPD (Table VI; available at www.jpeds.com). After adjustment for these factors, we found that IL-6 and IL-8 levels were associated with severity of BPD (P = .001 and .024, respectively) (Table V). For a 10-fold increase in IL-6 and IL-8, the risk of severe BPD vs all other categories was 71% and 77% higher, repectively; likewise, with the ordered logistic regression model the same OR reflects the risk of the combined mild, moderate, and severe BPD (any BPD) vs no BPD. To assess for biases associated with excluding the 28 deceased patients, we re-ran the analyses in Table V to include these patients (Table VII; available at www.jpeds.com) and obtained similar results with biomarker levels and severity of BPD.
Table VI.
Study demographics by BPD severity
| Characteristics | Population for which determination of BPD is possible (n = 411) | No BPD (n = 99) | Mild BPD (n = 83) | Moderate BPD (n = 131) | Severe BPD (n = 98) | P value |
|---|---|---|---|---|---|---|
| Maternal age (years) | 28.8 ± 6.5 [16–51] | 29.4 ± 6.9 [16–51] | 28.3 ± 6.0 [18–44] | 28.9 ± 6.5 [15–50] | 28.5 ± 6.7 [15–46] | .68 |
| Oligohydramnios (yes) | 9.3 | 12.1 | 8.4 | 6.1 | 11.2 | .39 |
| Intrauterine growth restriction (yes) | 18.7 | 20.2 | 12.1 | 16.0 | 26.5 | .07 |
| Prenatal steroids (yes) | 92.9 | 99.0 | 94.0 | 91.6 | 87.8 | .018 |
| Cesarean delivery (yes) | 62.0 | 64.6 | 54.2 | 67.9 | 58.2 | .17 |
| Sex (female) | 47.5 | 52.5 | 55.4 | 42.8 | 41.8 | .14 |
| Race/ethnicity (non-White) | 67.4 | 74.8 | 66.3 | 66.4 | 62.2 | .29 |
| Gestational age (weeks) | 27.7 ± 2.2 [23.1–31.9] | 30.0 ± 1.1 [26.7–31.9] | 27.4 ± 1.7 [23.9–30.9] | 27.0 ± 2.0 [23.1–31.9] | 26.3 ± 2.0 [23.3–31.3] | <.001 |
| Birthweight (g) | 974 ± 272 [420–1490] | 1229 ± 158 [810–1490] | 996 ± 222 [500–1450] | 922 ± 252 [470–1470] | 769 ± 217 [420–1480] | <.001 |
| RDS (yes) | 91.7 | 77.8 | 90.4 | 97.0 | 100.0% | <.001 |
Values are mean ± SD [range] or percent.
Table VII.
Predicted risk of BPD or death, including deceased patients
| Characteristics | Risk of BPD* (n = 439) | P value |
|---|---|---|
| VEGF | 0.98 ± 0.10 [0.80–1.20] | .83 |
| IL-6 | 1.71 ± 0.26 [1.27–2.30] | <.001 |
| IL-8 | 1.71 ± 0.41 [1.06–2.74] | .027 |
| IL-10 | 1.14 ± 0.17 [0.86–1.53] | .36 |
Values are OR ± SE [95% CI].
This analysis was performed to assess for bias from excluding the 28 deceased patients as was done in Table V. Ordered logistic regression with independent variable of biomarker (mean log-transformed levels from days 0 to 7 of life) and dependent categorical variable of BPD severity (0 = no BPD; 1 = mild BPD; 2 = moderate BPD; 3 = BPD) adjusted for gestational age, intrauterine growth restriction, birthweight, RDS, and prenatal steroids. For a 10-fold increase in biomarker level, the odds of severe BPD vs all other categories of BPD is the aOR; likewise, the same OR reflects the risk of the combined mild, moderate, and severe BPD (any BPD) vs no BPD. patients who were deceased before 36 weeks PMA were arbitrarily coded as having severe BPD.
Outpatient Respiratory Outcomes
Out of the 447 patients in the study population, 79 were subsequently enrolled in an outpatient pulmonary registry (Table VIII; available at www.jpeds.com). Of these patients, 74 had at least one questionnaire concerning outpatient respiratory outcomes before 3 years of age for a total of 152 forms. Potential confounders for outcomes included in the regression models were gestational age, severity of BPD, and age at form completion; regressions were clustered by patient to account for potentially more than one form per patient. Outcomes examined included acute care use (ie, emergency department visits and hospitalizations for respiratory symptoms, steroid courses, antibiotic courses for respiratory illnesses over the preceding 2 months) and chronic symptoms (ie, cough, congestion, wheezing, coughing with feeds, nighttime cough, use of short-acting beta-agonists at home, tolerance of physical activity over the past week). No outpatient respiratory outcomes were associated with biomarker levels at birth (Table IX; available at www.jpeds.com).
Table VIII.
Study demographics for outpatients
| Characteristics | Outpatient population (n = 74) |
|---|---|
| Maternal age (years) | 28.4 ± 6.4 [15–44] |
| Oligohydramnios (yes) | 13.5 |
| Intrauterine growth restriction (yes) | 20.3 |
| Prenatal steroids (yes) | 90.5 |
| Cesarean delivery (yes) | 58.1 |
| Sex (female) | 48.7 |
| Race/ethnicity (non-White) | 67.6 |
| Gestational age (weeks) | 26.2 ± 1.7 [23.4–30.0] |
| Birthweight (g) | 804 ± 228 [420–1450] |
| RDS (yes) | 94.6 |
| BPD | |
| None | 4.1 |
| Mild | 12.2 |
| Moderate | 35.1 |
| Severe | 48.7 |
| Mean age at questionnaire completion (averaged per patient; years) | 1.05 ± 0.52 [0.30–2.21] |
| Mean No. of questionnaires completed | 2.05 ± 1.29 [1–6] |
Values are mean ± SD [range] or percent.
Table IX.
Predicted risk for outpatient respiratory outcomes
| Outcomes | Mean log VEGF | Mean log IL-6 | Mean log IL-8 | Mean log IL-10 |
|---|---|---|---|---|
| Emergency deptartment visit (n = 148 forms) | 1.21 ± 0.43 | 0.61 ± 0.20 | 1.35 ± 1.00 | 0.64 ± 0.23 |
| P = .59* | P = .13* | P = .69* | P = .21* | |
| Hospital admission (n = 148 forms) | 0.63 ± 0.24 | 0.60 ± 0.23 | 1.69 ± 1.45 | 0.82 ± 0.32 |
| P = .22* | P = .18* | P = .54* | P = .61* | |
| Systemic steroid use (n = 150 forms) | 0.66 ± 0.20 | 0.69 ± 0.16 | 0.95 ± 0.69 | 0.84 ± 0.29 |
| P = .16 | P = .11 | P = .94 | P = .61 | |
| Antibiotic use (n = 135 forms) | 1.01 ± 0.35 | 0.58 ± 0.22 | 0.87 ± 0.52 | 0.73 ± 0.23 |
| P = .97* | P = .15* | P = .81* | P = .31* | |
| Cough/wheeze (n = 145 forms) | 1.39 ± 0.28 | 0.70 ± 0.16 | 0.64 ± 0.35 | 0.66 ± 0.14 |
| P = .10* | P = .13* | P = .42* | P = .05* | |
| Rescue β-agonist use (n = 146 forms) | 1.22 ± 0.27 | 1.25 ± 0.25 | 1.33 ± 0.69 | 0.71 ± 0.19 |
| P = .37 | P= .27 | P = .58 | P = .21 | |
| Activity limitations (n = 138 forms) | 1.49 ± 0.32 | 0.79 ± 0.17 | 0.58 ± 0.48 | 1.14 ± 0.26 |
| P = .07* | P = .27* | P = .51* | P = .56* | |
| Nighttime symptoms (n = 143 forms) | 1.18 ± 0.33 | 0.42 ± 0.12 | 0.29 ± 0.21 | 0.65 ± 0.21 |
| P = .55* | P = .002* | P = .09* | P = .18* |
Values are OR ± SE clustered (by patient) logistic regression with independent variable of biomarker (mean log-transformed levels from days 0 to 7 of life) and dependent variable of dichotomous outcome adjusted for gestational age, severity of BPD (categorical dummy variable), and age at form completion. The threshold P value corrected for multiple testing (32 tests) is 0.0016.
Collinearity identified in regression.
Discussion
We identified several associations between specific cytokines and VEGF and RDS and BPD severity in a gradated fashion. Several of these associations persisted after adjustment for confounding factors. However, we did not observe any associations between perinatal cytokines and respiratory outcomes after initial hospital discharge in children up to 3 years of age, indicating that the plasma levels of these proteins (VEGF, IL-6, IL-8, and IL-10) during the first week of life may not be able to predict respiratory outcomes in children with BPD in the outpatient setting.
We observed that the levels of VEGF, IL-6, and IL-10 obtained within the first 24 hours of life were associated with RDS and that, after adjustment for relevant factors, lower levels of VEGF were associated with an increased frequency of RDS. Tsao et al using cord blood also reported that lower levels of VEGF were associated with higher rates of RDS.19 Lower levels of VEGF tend to be seen in infants with RDS across several human studies.20 However, potential treatment with VEGF may be associated with side effects, including pulmonary edema and/or hemorrhage, and possible implications for retinopathy of prematurity.20–22 Alternatively, lower levels of VEGF may be a biomarker for poor fetal lung growth before delivery, thus predisposing the preterm infant to the development of RDS. Support for this finding includes a preclinical study in mice that found that prenatal hypoxia impaired lung growth, decreased expression of pulmonary hypoxia-inducible factor-1α and hypoxia-inducible factor-2α expression and VEGF protein level in fetal lungs.23 Similar to D’Angio et al, we did not observe associations between IL-6, IL-8, or IL-10 and RDS.11
Although BPD is defined by oxygen use and respiratory support, it has a complex pathophysiology, incorporating elements of alveolar, airway, and pulmonary vascular disease. In our analysis, we found that IL-6 and IL-8 levels were associated with BPD at 36 weeks. Three prior studies have observed that elevated levels of IL-6, IL-8, and IL-10 were associated with BPD, although 2 studies did not.10–14 Supporting our observation, we also observed that the raw levels of Il-6 and IL-8 were associated with BPD severity in a gradated fashion as well. The proinflammatory role of IL-6 is well-established in the development of many diseases through its pleotropic effects, including promoting chronic inflammation through T-cell differentiation and its proangiogenic effects found in many malignancies.24,25 IL-8 has been found to be associated with neutrophilic chemotaxis and phagocytosis and, similar to IL-6, has proangiogenic properties in a tumor environment.26 Elevation of IL-6 and IL-8 may have relevance for disease progression toward a BPD phenotype.
One of the novel aspects of our study is follow-up into early childhood on a subgroup of preterm infants with established respiratory disease. In our study, we did not observe any associations between the selected biomarkers and outpatient acute or chronic respiratory outcomes before 3 years of age, despite seeing associations with respiratory outcomes at earlier timepoints. However, Hagman et al reported that higher levels of IL-6 (although not IL-8 or IL-10) at birth in 53 preterm infants born in Sweden were associated with obstructive lung disease as measured by pulmonary function testing (forced expiratory volume in 1 second and forced mid-expiratory flow) at 12 years of age.13 Although it is possible that our questionnaire is subject to more misclassification bias than quantitative pulmonary function tests, we would speculate that accumulation of other environmental exposures once outside the more stable environment of the NICU may dilute the predictive powers of biomarkers obtained within the first week of life in our population. Within our outpatient population, we have previously observed associations with respiratory outcomes from such environmental exposures such as socioeconomic factors, indoor air pollution, road proximity, breast milk, tobacco smoke exposure, and daycare.27–32 In addition to environmental factors that may alter the course of preterm respiratory disease, trajectories may also be influenced by genetic factors. Several prior twin studies have suggested a heritable component for BPD.33–35
The limitations of our study include only a small subset of the study population with outpatient data, which may be biased toward patients with more severe respiratory disease because outpatient data were collected from a pulmonary subspecialty clinic. Our overall study population was recruited from a single urban tertiary care center, and thus these results may not be as generalizable to other populations. Another important limitation is that RDS was assessed via discharge documentation, which may lack sensitivity for this outcome. We also assessed biomarkers through blood specimens, and levels are not lung specific.4 Natural fluctuations during the first week of life and gestational age may affect our results, although we adjusted for gestational age and averaged biomarker measurements for relevant time periods.36 Results in our study may differ from other published studies owing to the frequency of specific confounders that we examined; for example, we do not have data on maternal chorioamnionitis or early-onset sepsis. Additionally, our population may have higher or lower frequencies of specific single nucleotide polymorphisms associated with biomarker levels; some prior single nucleotide polymorphism studies have identified associations between BPD and VEGF single nucleotide polymorphisms.37
Even though we found that selected biomarkers are associated with respiratory outcomes for preterm infants within the NICU setting, they are potentially less useful outside the NICU setting. Given that we may be underpowered to examine outpatient outcomes, future studies could consider examining the role of biomarkers in outpatient respiratory outcomes in a larger population, or assessing the utility of biomarkers obtained at a later timepoint in the NICU course (eg, at 36 weeks corrected age or at discharge).
Acknowledgments
The authors thank the families who participated in this study.
Supported by National Institutes of Health National Institute of Child Health and Human Development R01HD086058 [to A.E., F.N., and E.G.], National Institute of Neurological Disorders and Stroke KO8NS096115 [to R.C.-V.], and the Thomas-Wilson Foundation [to R.C.-V.].
Glossary
- BPD
Bronchopulmonary dysplasia
- IL
Interleukin
- LLOD
Lower limits of detection
- NICU
Neonatal intensive care unit
- RDS
Respiratory distress syndrome
- ULOD
Upper limits of detection
- VEGF
Vascular endothelial growth factor
Footnotes
The authors declare no conflicts of interest.
References
- 1.Collaco JM, McGrath-Morrow SA. Respiratory phenotypes for preterm infants, children, and adults: bronchopulmonary dysplasia and more. Ann Am Thorac Soc 2018;15:530–8. [DOI] [PubMed] [Google Scholar]
- 2.Chan JY, Stern DA, Guerra S, Wright AL, Morgan WJ, Martinez FD. Pneumonia in childhood and impaired lung function in adults: a longitudinal study. Pediatrics 2015;135:607–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Martinez FD. Early-life origins of chronic obstructive pulmonary disease. N Engl J Med 2016;375:871–8. [DOI] [PubMed] [Google Scholar]
- 4.Lal CV, Ambalavanan N. Biomarkers, early diagnosis, and clinical predictors of bronchopulmonary dysplasia. Clin Perinatol 2015;42:739–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Fujinaga H, Baker CD, Ryan SL, Markham NE, Seedorf GJ, Balasubramaniam V, et al. Hyperoxia disrupts vascular endothelial growth factor-nitric oxide signaling and decreases growth of endothelial colony-forming cells from preterm infants. Am J Physiol Lung Cell Mol Physiol 2009;297:L1160–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Abman SH. Impaired vascular endothelial growth factor signaling in the pathogenesis of neonatal pulmonary vascular disease. Adv Exp Med Biol 2010;661:323–35. [DOI] [PubMed] [Google Scholar]
- 7.McGrath-Morrow S, Cho C, Molls R, Burne-Taney M, Haas M, Hicklin DJ, et al. VEGF receptor 2 blockade leads to renal cyst formation in mice. Kidney Int 2006;69:1741–8. [DOI] [PubMed] [Google Scholar]
- 8.You Y, Guo C, Zhang H, Deng S, Tang J, Xu L, et al. Effect of intranasal instillation of lipopolysaccharide on lung development and its related mechanism in newborn mice. J Interferon Cytokine Res 2019;39:684–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Sahni M, Bhandari V. Recent advances in understanding and management of bronchopulmonary dysplasia. F1000Res 2020;9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Ambalavanan N, Carlo WA, D’Angio CT, McDonald SA, Das A, Schendel D, et al. Cytokines associated with bronchopulmonary dysplasia or death in extremely low birth weight infants. Pediatrics 2009;123:1132–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.D’Angio CT, Ambalavanan N, Carlo WA, McDonald SA, Skogstrand K, Hougaard DM, et al. Blood cytokine profiles associated with distinct patterns of bronchopulmonary dysplasia among extremely low birth weight infants. J Pediatr 2016;174:45–51 e5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Paananen R, Husa AK, Vuolteenaho R, Herva R, Kaukola T, Hallman M. Blood cytokines during the perinatal period in very preterm infants: relationship of inflammatory response and bronchopulmonary dysplasia. J Pediatr 2009;154:39–43 e3. [DOI] [PubMed] [Google Scholar]
- 13.Hagman C, Bjorklund LJ, Bjermer L, Hansen-Pupp I, Tufvesson E. Perinatal inflammation relates to early respiratory morbidity and lung function at 12 years of age in children born very preterm. Acta Paediatr 2021;110:2084–92. [DOI] [PubMed] [Google Scholar]
- 14.Sahni M, Yeboah B, Das P, Shah D, Ponnalagu D, Singh H, et al. Novel biomarkers of bronchopulmonary dysplasia and bronchopulmonary dysplasia-associated pulmonary hypertension. J Perinatol 2020;40: 1634–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Jobe AH, Bancalari E. Bronchopulmonary dysplasia. Am J Respirat Crit Care Med 2001;163:1723–9. [DOI] [PubMed] [Google Scholar]
- 16.Higgins RD, Jobe AH, Koso-Thomas M, Bancalari E, Viscardi RM, Hartert TV, et al. Bronchopulmonary dysplasia: executive summary of a workshop. J Pediatr 2018;197:300–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Hendricks-Munoz KD, Xu J, Voynow JA. Tracheal aspirate VEGF and sphingolipid metabolites in the preterm infant with later development of bronchopulmonary dysplasia. Pediatr Pulmonol 2018;53: 1046–52. [DOI] [PubMed] [Google Scholar]
- 18.Dietrick B, Molloy E, Massaro AN, Strickland T, Zhu J, Slevin M, et al. Plasma and CSF candidate biomarkers of neonatal encephalopathy severity and neurodevelopmental outcomes. J Pediatr 2020;226: 71–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Tsao PN, Wei SC, Chou HC, Su YN, Chen CY, Hsieh FJ, et al. Vascular endothelial growth factor in preterm infants with respiratory distress syndrome. Pediatr Pulmonol 2005;39:461–5. [DOI] [PubMed] [Google Scholar]
- 20.Meller S, Bhandari V. VEGF levels in humans and animal models with RDS and BPD: temporal relationships. Exp Lung Res 2012;38: 192–203. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Kunig AM, Balasubramaniam V, Markham NE, Seedorf G, Gien J, Abman SH. Recombinant human VEGF treatment transiently increases lung edema but enhances lung structure after neonatal hyperoxia. Am J Physiol Lung Cell Molec Physiol 2006;291:L1068–78. [DOI] [PubMed] [Google Scholar]
- 22.Sankar MJ, Sankar J, Chandra P. Anti-vascular endothelial growth factor (VEGF) drugs for treatment of retinopathy of prematurity. Cochrane Database Syst Rev 2018;1:CD009734. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Tsao PN, Wei SC. Prenatal hypoxia downregulates the expression of pulmonary vascular endothelial growth factor and its receptors in fetal mice. Neonatology 2013;103:300–7. [DOI] [PubMed] [Google Scholar]
- 24.Kumari N, Dwarakanath BS, Das A, Bhatt AN. Role of interleukin-6 in cancer progression and therapeutic resistance. Tumour Biol 2016;37: 11553–72. [DOI] [PubMed] [Google Scholar]
- 25.Tanaka T, Narazaki M, Kishimoto T. IL-6 in inflammation, immunity, and disease. Cold Spring Harb Perspect Biol 2014;6:a016295. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Alfaro C, Sanmamed MF, Rodríguez-Ruiz ME, Teijeira Á, Oñate C, González Á, et al. Interleukin-8 in cancer pathogenesis, treatment and follow-up. Cancer Treat Rev 2017;60:24–31. [DOI] [PubMed] [Google Scholar]
- 27.Collaco JM, Choi SJ, Riekert KA, Eakin MN, McGrath-Morrow SA, Okelo SO. Socio-economic factors and outcomes in chronic lung disease of prematurity. Pediatr Pulmonol 2011;46:709–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Rice JL, McGrath-Morrow SA, Collaco JM. Indoor air pollution sources and respiratory symptoms in bronchopulmonary dysplasia. J Pediatr 2020;222:85–90 e2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Collaco JM, Morrow M, Rice JL, McGrath-Morrow SA. Impact of road proximity on infants and children with bronchopulmonary dysplasia. Pediatr Pulmonol 2020;55:369–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Kim LY, McGrath-Morrow SA, Collaco JM. Impact of breast milk on respiratory outcomes in infants with bronchopulmonary dysplasia. Pediatr Pulmonol 2019;54:313–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Collaco JM, Aherrera AD, Breysse PN, Winickoff JP, Klein JD, McGrath-Morrow SA. Hair nicotine levels in children with bronchopulmonary dysplasia. Pediatrics 2015;135:e678–86. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.McGrath-Morrow SA, Lee G, Stewart BH, McGinley BM, Lefton-Greif MA, Okelo SO, et al. Day care increases the risk of respiratory morbidity in chronic lung disease of prematurity. Pediatrics 2010;126:632–7. [DOI] [PubMed] [Google Scholar]
- 33.Lavoie PM, Pham C, Jang KL. Heritability of bronchopulmonary dysplasia, defined according to the consensus statement of the national institutes of health. Pediatrics 2008;122:479–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Bhandari V, Bizzarro MJ, Shetty A, Zhong X, Page GP, Zhang H, et al. Familial and genetic susceptibility to major neonatal morbidities in preterm twins. Pediatrics 2006;117:1901–6. [DOI] [PubMed] [Google Scholar]
- 35.Parker RA, Lindstrom DP, Cotton RB. Evidence from twin study implies possible genetic susceptibility to bronchopulmonary dysplasia. Semin Perinatol 1996;20:206–9. [DOI] [PubMed] [Google Scholar]
- 36.Lusyati S, Hulzebos CV, Zandvoort J, Sauer PJ. Levels of 25 cytokines in the first seven days of life in newborn infants. BMC Res Notes 2013;6:547. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Piersigilli F, Bhandari V. Biomarkers in neonatology: the new “omics” of bronchopulmonary dysplasia. J Matern Fetal Neonatal Med 2016;29: 1758–64. [DOI] [PubMed] [Google Scholar]