Thursday 16 June 2011
RSV bronchiolitis in healthy term infants: prediction and pathogenesis
Promotor: Prof.dr J.L.L. Kimpen and prof.dr G.H.A. Visser
Defence: 16 June 2011
Respiratory syncytial virus (RSV) is the most important cause of bronchiolitis in infants. Annually, an estimated total of 34 million children under the age of 5 worldwide develop severe RSV infection, with 3.4 million being hospitalized. In the United States more than 2 million children under the age of 5 develop medically attended RSV infection each year, of which 97% are treated as outpatients. In half of the infants with RSV bronchiolitis recurrent wheeze during preschool years develops resulting in health-related quality of life (HRQoL).
Of all children that are hospitalized because of RSV bronchiolitis, approximately 70% have no underlying disorder. Moreover, most children that develop RSV bronchiolitis in general have no known risk factors, such as congenital heart disease, chronic lung disease, or premature birth. Establishing new predictive factors and a prediction rule for RSV bronchiolitis in healthy term infants could help to target preventive and early therapeutic interventions in future. Moreover, the etiology of RSV bronchiolitis in healthy term infants is largely unravelled and new insights may ultimately contribute to the development of new strategies to prevent or treat RSV bronchiolitis.
This thesis describes a series of studies on the prediction and the pathogenesis of RSV bronchiolitis in healthy term infants, all carried out in the Netherlands Amniotic Fluid (NAF) cohort, a healthy term birth cohort based in the University Medical Center Utrecht and the Diakonessen Hospital (both Utrecht, The Netherlands).
In Chapter 1, RSV bronchiolitis in healthy term infants is comprehensively introduced. The majority of medically attended RSV infections occur in children above the age of six months and do not require hospital admission. Nevertheless, the socioeconomic burden of RSV infections is substantial. The general aims of this thesis and the specific research questions of the studies on RSV prediction and pathogenesis in this thesis are discussed.
In Chapter 2, novel predictive factors for RSV bronchiolitis in the community and a simple clinical prediction rule are presented. In the NAF birth cohort study, 298 healthy term newborns were followed throughout the first year of life. Parents kept daily logs and collected nose-throat swabs during respiratory tract infections. The primary outcome was RSV lower respiratory tract infection (LRTI), which was defined as the combination of a positive RSV PCR and acute wheeze or moderate to severe cough. Of the participating children, 42 (14%) developed RSV LRTI. Independent predictors for RSV LRTI were: day-care / siblings, high parental education, birth weight >4 kg, and birth in April-September. The discriminative power of the prediction model was moderately high (ROC AUC 0.72). A clinical prediction rule was derived, with possible scores ranging from 0 to 5 points. The absolute risk of RSV LRTI was 3% in children with a score ≤ 2 and 32% for the children with score 5. Furthermore, 62% of the children with RSV LRTI suffered from wheezing during the first year of life versus 36% in those without. We concluded that a simple clinical prediction rule can identify healthy newborns at risk of RSV LRTI. Physicians can differentiate between children with high and low risk of RSV LRTI, and subsequently target preventive and monitoring strategies at high risk children.
In Chapter 3, the association between plasma concentrations of 25-OH vitamin D at birth and the subsequent risk of RSV LRTI was assessed. In the NAF birth cohort study, cord blood concentrations of 25-OH vitamin D were related to risk of RSV LRTI in the first year of life. Of 156 healthy term newborns, 18 (12%) developed RSV LRTI. The mean cord blood concentration of 25-OH vitamin D was 82 nmol/L. Overall, 27% of newborns had 25-OH vitamin D concentrations <50 nmol/L. Maternal vitamin D3 supplementation during pregnancy was positively associated with cord blood 25-OH vitamin D. Concentrations of 25-OH vitamin D were lower in newborns who developed RSV LRTI compared to those who did not (65 versus 84 nmol/L). Newborns with 25-OH vitamin D <50 nmol/L had sixfold increased risk of RSV LRTI in the first year of life compared to those with 25-OH vitamin D 75 nmol/L. We concluded that vitamin D deficiency at birth is associated with increased risk of RSV LRTI in the first year of life. Intensified routine vitamin D supplementation during pregnancy may be a useful strategy to prevent RSV LRTI during infancy.
In Chapter 4, we hypothesized that disease severity is correlated with viral load in primary RSV infection in the community. Eighty-two healthy term infants of the NAF birth cohort were included at signs of their first respiratory tract infection. Nasopharyngeal aspirate was obtained within 48-96 hours and disease severity was assessed with a previously published severity scoring model. PCR was applied to test the aspirates in a semi-quantitative way for the presence of ten respiratory pathogens. In case of multiple infection, the pathogen with the highest load was defined as the primary pathogen. Median age at first respiratory tract infection was 3 months. Pathogens were detected in 77 infants (94%); more than one pathogen was detected in 35 infants (43%). RSV was present in aspirates of 30 infants; in 16 aspirates RSV was the primary pathogen. A negative correlation between RSV CT-value and disease severity was found in all RSV cases (ρ= -0.52) and in cases with RSV as the primary pathogen (ρ= -0.54). This is the first report on viral loads in previously healthy infants with primary RSV infection in the community. We concluded that disease severity correlated positively with viral load during primary RSV infection.
In Chapter 5, we investigated the sensitivity of nasal swabs and nasopharyngeal aspirates and assessed whether viral load played a role in the sensitivity of either sampling method. Healthy term infants of the NAF birth cohort were included at signs of first respiratory tract infection. Paired nasopharyngeal aspirates and nasal swabs were collected and real-time PCR was carried out as before mentioned. Paired samples were collected in 98 infants, with at least one respiratory pathogen identified in 94 cases. Rhinovirus (n=67) and RSV (n=39) were most frequently detected. The sensitivity of the nasal swab was lower than the nasopharyngeal aspirate, in particular for RSV (51% versus 100%) and rhinovirus (75% versus 97%). Sensitivity of the nasal swab was strongly determined by the CT-value. Sensitivity of the swab for RSV, but not rhinovirus, was 100% in children with severe symptoms (score ≥11). We concluded that the nasal swab may be used for community based studies on RSV infection, though its sensitivity is lower than of the aspirate, in particular for detection of mild cases.
In Chapter 6, we hypothesized that intra-uterine inflammation at term is associated with spontaneous onset of labor. Within the NAF birth cohort, a cross-sectional study of 375 spontaneous onset of labor term vaginal deliveries and elective caesarean sections (CS), without signs of labor, was carried out. Placentas and amniotic fluid samples were collected during labor and / or at delivery. Histological signs of placenta inflammation were determined. Amniotic fluid pro-inflammatory cytokine concentrations were measured using ELISA. In term vaginal deliveries, more signs of intra-uterine inflammation were found than in elective CS: the prevalence of chorioamnionitis was higher (18 versus 4%) and the amniotic fluid concentration of interleukin-6 (IL-6) was higher (3.1 versus 0.4 ng/mL). Similar results were obtained for IL-8 (10.9 versus 1.0 ng/mL) and the proportion of deliveries with detectable amniotic fluid tumor necrosis factor-α (TNF-α, 50 versus 4%). We concluded that spontaneous term delivery is characterized by histological signs of placenta inflammation and increased amniotic fluid pro-inflammatory cytokines.
In Chapter 7, we determined whether high amniotic fluid IL-8 and TNF-α protect against RSV bronchiolitis in healthy term infants, using the NAF birth cohort. In case of medical attention for respiratory symptoms during the first year of life, a nose-throat swab was taken for RSV PCR. Physician-attended RSV infection was observed in 27 (9.3%) of 292. Amniotic fluid concentrations of IL-8 were higher in children without physician-attended RSV infection than in children with physician-attended RSV infection (11.1 versus 5.5 ng/mL). Similarly, in children without physician-attended RSV the proportion of detectable amniotic fluid TNF-α was higher (60% versus 30%). Among children with physician-attended RSV infection, amniotic fluid IL-8 was inversely correlated to the number of wheezing days during the first year of life (ρ=-0.38). We concluded that high concentrations of amniotic fluid IL-8 and TNF-α are associated with low risk of RSV bronchiolitis in healthy infants. We hypothesized that direct exposure of fetal lungs to pro-inflammatory signals induces local protection against viral infection during infancy.
In Chapter 8, intra-uterine immune activation was studied as a possible determinant of newborn airway function. Leukocyte-associated Ig-like receptor-1 (LAIR-1) is a membrane-bound collagen receptor that increases the threshold for activating signals on several immune cells. The secreted form of LAIR-1 (sLAIR-1) is considered a distinct marker of general immune activation. We hypothesized that high sLAIR-1 in amniotic fluid is associated with normal newborn airway function and low infant airway morbidity. In the NAF birth cohort, 152 healthy term newborns successfully underwent lung function measurement. At age 1 month, airway compliance and resistance were assessed with the single occlusion technique. sLAIR-1 was detected in all amniotic fluid samples. Airway compliance and amniotic fluid concentration of sLAIR-1 were positively correlated (ρ=0.29). This correlation was independent from sex or maternal antepartum smoking. Amniotic fluid sLAIR-1 was lower in children who wheezed at ages 6 and 9 months. We concluded that high amniotic fluid sLAIR-1 concentration provide support of a strong intra-uterine immune activation during normal delivery. An association between sLAIR-1 in amniotic fluid and newborn airway compliance suggests a beneficial effect of intra-uterine immune activation on neonatal lung function.
Finally, in Chapter 9, a general discussion on RSV bronchiolitis in healthy term infants is given. First, the prediction of RSV bronchiolitis in healthy term infants is discussed. Abundant exposure to the virus is the most important predictive factor for RSV bronchiolitis in healthy term infants. Furthermore, birth approximately six months before the peak of the RSV epidemic also increased the risk, probably due to diminished maternal antibodies during the epidemic. Additionally, vitamin D deficiency at birth is associated with increased risk of RSV bronchiolitis. Upon external validation, the prediction rule for RSV bronchiolitis in healthy term infants could be used to target preventive and therapeutic strategies. Second, the pathogenesis of RSV bronchiolitis in healthy term infants is discussed. Viral load determines disease severity of RSV infections in the community, as in children hospitalized with RSV bronchiolitis. Remarkably, the absence of high amniotic fluid levels of pro-inflammatory cytokines confers an increased risk of RSV bronchiolitis. Possibly, this finding is mediated by an impaired maturation of the innate immune system of the airways. Future research should focus on the origins and effects of intra-uterine inflammation during term physiologic parturition, in order to understand whether exposure to high concentrations of pro-inflammatory cytokines is causally linked to infant lung function and respiratory disease.