Clinical Prediction Rule for RSV Bronchiolitis in Healthy Newborns: Prognostic Birth Cohort Study
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
ARTICLES
Clinical Prediction Rule for RSV Bronchiolitis in
Healthy Newborns: Prognostic Birth Cohort Study
AUTHORS: Michiel L. Houben, MD,a Louis Bont, MD, PhD,a WHAT’S KNOWN ON THIS SUBJECT: Hospitalized respiratory
Berry Wilbrink, PhD,b Mirjam E. Belderbos, MD,a Jan L. L. syncytial virus (RSV) lower respiratory tract infection (LRTI) can
Kimpen, MD, PhD,a Gerard H. A. Visser, MD, PhD,c and be predicted by using host and environmental factors. The impact
Maroeska M. Rovers, PhDd of outpatient-treated RSV LRTI includes increased number of
aDepartment of Pediatrics, Wilhelmina Children’s Hospital, physician visits, drug prescriptions, and parents’ missed work
cDepartment of Obstetrics and Gynecology, and dJulius Center days.
for Health Sciences and Primary Care, University Medical Center
Utrecht, Utrecht, Netherlands; and bLaboratory of Infectious
Diseases and Perinatal Screening, National Institute of Public WHAT THIS STUDY ADDS: A simple prediction rule can identify
Health and the Environment, Bilthoven, Netherlands infants at risk of outpatient-treated RSV LRTI. The absolute risks
KEY WORDS of RSV LRTI range from 3% for children with the lowest prediction
birth cohort study, respiratory syncytial virus, lower respiratory rule score to 32% for children with all predictive factors.
tract infection, health-related quality of life, postbronchiolitis
wheeze, risk stratification
ABBREVIATIONS
AUC—area under the curve
GP—general practitioner
HRQoL—health-related quality of life
abstract +
LRTI—lower respiratory tract infection OBJECTIVE: Our goal was to determine predictors of respiratory syn-
OR—odds ratio
PCR—polymerase chain reaction
cytial virus (RSV) lower respiratory tract infection (LRTI) among
ROC—receiver operating characteristic healthy newborns.
RSV—respiratory syncytial virus
METHODS: In this prospective birth cohort study, 298 healthy term
www.pediatrics.org/cgi/doi/10.1542/peds.2010-0581 newborns born in 2 large hospitals in the Netherlands were monitored
doi:10.1542/peds.2010-0581 throughout the first year of life. Parents kept daily logs and collected
Accepted for publication Oct 8, 2010 nose/throat swabs during respiratory tract infections. The primary
Address correspondence to Louis Bont, MD, PhD, University outcome was RSV LRTI, which was defined on the basis of the combina-
Medical Center Utrecht, Department of Pediatrics, Lundlaan 6,
tion of positive RSV polymerase chain reaction results and acute
3584 EA Utrecht, Netherlands. E-mail: l.j.bont@umcutrecht.nl
wheeze or moderate/severe cough.
PEDIATRICS (ISSN Numbers: Print, 0031-4005; Online, 1098-4275).
Copyright © 2011 by the American Academy of Pediatrics
RESULTS: Of the 298 children, 42 (14%) developed RSV LRTI. Indepen-
dent predictors for RSV LRTI were day care attendance and/or siblings,
FINANCIAL DISCLOSURE: Dr Bont received research funding and
speaker’s fees from Abbott International; the other authors have high parental education level, birth weight of ⬎4 kg, and birth in April
indicated they have no financial relationships relevant to this to September. The area under the receiver operating characteristic
article to disclose.
curve was 0.72 (95% confidence interval: 0.64 – 0.80). We derived a
clinical prediction rule; possible scores ranged from 0 to 5 points. The
absolute risk of RSV LRTI was 3% for children with scores of ⱕ2 (20% of
all children) and 32% for children with all 4 factors (scores of 5; 8% of
all children). Furthermore, 62% of the children with RSV LRTI experi-
enced wheezing during the first year of life, compared with 36% of the
children without RSV LRTI.
CONCLUSIONS: A simple clinical prediction rule identifies healthy new-
borns at risk of RSV LRTI. Physicians can differentiate between children
with high and low risks of RSV LRTI and subsequently can target pre-
ventive and monitoring strategies toward children at high risk.
Pediatrics 2011;127:35–41
PEDIATRICS Volume 127, Number 1, January 2011 35The proportion of all children in their senhuis [Netherlands]) participated in during any period in the first year of
first year of life with medically at- this prospective birth cohort study. life. The presence of siblings in the
tended respiratory syncytial virus Children who were born after 37 household of the child was defined as
(RSV) infections in the United States is weeks of gestation (term) after an un- ⱖ1 sibling younger than 18 years liv-
estimated to be 44%.1 The majority of complicated pregnancy were eligible ing ⱖ3 days per week in the same
these children (95%) are treated as to participate. Newborns with major house. A composite variable of day
outpatients by general practitioners congenital anomalies and newborns care and/or siblings was created to
(GPs) or at the emergency depart- whose parents had limited Dutch lan- limit the number of potential predic-
ment.1 Therefore, from a socioeco- guage skills were excluded. Between tive factors, because of the relatively
nomic point of view, outpatient-treated January 2006 and December 2008, small sample size. Parental education
RSV infections have a large impact, in- 1080 newborns were eligible and the level was dichotomized by using the ar-
cluding emergency department and of- parents of 341 (32%) agreed to partic- bitrary cutoff level of a bachelor’s de-
fice visits, drug prescriptions, and par- ipate and gave written informed con- gree for ⱖ1 parent. Because maternal
ents’ missed work days.2 Moreover, sent. The most frequent reason for anti-RSV antibodies may protect in-
RSV lower respiratory tract infection nonparticipation was reluctance of fants against RSV disease in the com-
(LRTI) is associated with subsequent parents to perform daily follow-up munity in their first months of life,
recurrent wheeze for ⬃40% of pa- measurements according to the study being born within 6 months before
tients, leading to reduced health- protocol. Baseline characteristics of the start of the RSV season (April
related quality of life (HRQoL).3–5 nonparticipating children and their through September) was used as a
Identifying newborns who will develop parents were similar to the character- potentially predictive variable.25
RSV LRTI is important, because simple istics of participating subjects (data
lifestyle changes, such as intensified not shown). Of the 341 included chil- Outcomes
hand hygiene, can prevent RSV infec- dren, 298 (87%) had no missing values.
The primary outcome was RSV LRTI,
tions.6–9 In addition, current and future The study protocol was approved by
which was defined as the presence of
medical preventive measures may be the institutional review boards of the 2
an LRTI and the presence of RSV RNA.
used to target individuals at high participating hospitals.
Parents were instructed to record
risk.10,11 Known risk factors for the oc- daily respiratory symptoms, including
Predictive Factors
currence of RSV LRTI are preterm wheeze and cough, in a log.26 Episodes
birth, young age, male gender, heart The presence or absence of risk fac-
in the log were defined to represent an
and lung disease, Down syndrome, ab- tors for RSV LRTI was assessed by us-
LRTI by using strict predefined criteria,
sence or short duration of breastfeed- ing data from the hospital delivery files
that is, moderate or severe cough or
ing, presence of siblings, day care at- (gender, gestational age, birth weight,
wheeze of any severity lasting for ⱖ2
tendance, and exposure to tobacco and month of birth) or from standard-
days. A nose/throat swab sample was
smoke.12–20 ized questionnaires completed at 1
obtained by the parents at the start of
month and 1 year of age. Gestational
To date, clinical prediction models for every respiratory episode and subse-
age was dichotomized by using an ar-
RSV have been developed only with re- quently was sent to the researchers in
bitrary cutoff value of 40.0 weeks. Birth
spect to hospitalization among pre- a single vial containing 2 mL of viral
weight was dichotomized by using an
term infants.21–23 A clinical prediction transport medium. The samples were
arbitrary cutoff value of 4 kg.24 Breast-
model for outpatient-treated RSV LRTI frozen at ⫺80°C until polymerase
feeding was defined as being given
among term children does not yet ex- chain reaction (PCR) assays were per-
mother’s milk exclusively (without ad-
ist. Therefore, the objective of this formed. The presence of RSV A or B RNA
ditional formula feeding) beyond the
study was to develop a clinical predic- was determined by using real-time
age of 1 month. Parental atopy was de-
tion rule to identify healthy term new- PCR assays.27
fined as the presence of any atopic di-
borns at high risk of RSV LRTI in the
agnosis (asthma, eczema, or hay fe- A secondary outcome was GP-attended
first year of life.
ver) made by a physician for 1 or both RSV infection, which was defined as
METHODS parents. Exposure to maternal tobacco the occurrence of a respiratory epi-
smoke was defined as maternal smok- sode with GP attendance and the pres-
Population ing of ⱖ1 cigarette per day at the age ence of RSV RNA. To study the burden of
Two large urban hospitals (University of 1 month. Day care attendance was RSV LRTI episodes, we also examined
Medical Center Utrecht and Diakones- defined as attendance of any day care wheezing during the first year of life
36 HOUBEN et alARTICLES
and HRQoL (measured with the TNO-AZL
Preschool Children Quality of Life ques-
tionnaire [TAPQoL]) for children with
and without RSV LRTI (secondary out-
comes).28–30 Data on wheezing during
the first year of life were derived from
the logs.
Statistical Analyses
The association between each prog-
nostic factor and the presence or ab-
sence of RSV LRTI was examined with
univariate logistic regression analy-
ses. Predictors that were associated
with the outcome in univariate analy-
ses (P ⬍ .15) were included in multi-
variate logistic regression analyses.
The model was reduced through exclu- FIGURE 1
Distributions of month of birth for children with (A) and without (B) RSV LRTI(s) and of month of RSV
sion of predictors with P values of LRTI(s) (C).
⬎.10. The predictive accuracy of the
model was estimated on the basis of
its reliability (goodness of fit) by using strap sample. This yielded a shrinkage with wheeze each month, respiratory
Hosmer-Lemeshow tests.31,32 The mod- factor for the regression coefficients symptoms, and HRQoL between chil-
el’s ability to discriminate between and the ROC AUC.33 dren with and children without RSV
children with and without RSV LRTI was To obtain a prediction rule that is eas- LRTI in the first year of life. All analyses
estimated as the area under the curve ily applicable in clinical practice, the were performed with SPSS 15 (SPSS
(AUC) for the receiver operating char- adjusted regression coefficients of the Inc, Chicago, IL).
acteristic (ROC) curve for the model. model were divided by the lowest coef-
The ROC curve is a plot of the true- ficient and rounded to the nearest in- RESULTS
positive rate (sensitivity) versus the teger. Scores for each individual pa- Of the 298 participating children, 42
false-positive rate (1 ⫺ specificity) tient were obtained by assigning (14%) developed RSV LRTIs during
evaluated at consecutive cutoff points points for each variable and adding their first year of life. One child devel-
for the predicted probability. The AUC the results. Patients were classified oped 2 separate RSV LRTIs within the
provides a quantitative summary of according to their risk scores and the same season. The median age at the
the discriminative ability of a predic- number of children developing or not time of RSV LRTI was 6 months (inter-
tive model. A useless predictive model, developing RSV LRTI, and correspond- quartile range: 4 – 8 months) (Fig 1).
such as a coin flip, would yield an AUC ing positive and negative predictive Twenty children (48%) were boys. Of
of 0.5. When the AUC is 1.0, the model values were calculated. the 42 children with RSV LRTI, 27 (64%)
discriminates perfectly between sub- visited a GP and 3 (2, 6, and 8 months of
To test the robustness of the model,
jects who do and subjects who do not age) were hospitalized. Although RSV A
sensitivity analyses were conducted by
develop a prognostic outcome.33 using the alternative outcome of GP- (25 of 42 cases) and RSV B (17 of 42
Prediction models derived with multi- attended RSV infection and by using al- cases) were detected separately in
variate regression analyses are known ternative predictive factors (eg, day PCR assays, the clinical outcomes of
for overestimated regression coeffi- care and siblings as separate vari- children with LRTI attributable to RSV A
cients, which result in too-extreme ables and duration of breastfeeding, and RSV B were comparable (data not
predictions when applied in new cas- intensity of maternal smoking, and du- shown).
es.33 Therefore, we validated our mod- ration of day care attendance as con- Results of univariate and multivariate
els internally with bootstrapping tech- tinuous variables). The clinical rele- logistic regression analyses are pre-
niques in which the entire modeling vance of the model was studied by sented in Table 1. The final reduced re-
process was repeated with each boot- comparing the proportions of children gression model included 4 indepen-
PEDIATRICS Volume 127, Number 1, January 2011 37TABLE 1 Univariable and Multivariable Analyses of Predictors of RSV LRTI
Characteristic n (%) Univariate Analyses Multivariate Analyses Points
(Final Model) for
Rule
RSV LRTI No RSV LRTI OR (95% CI) P OR (95% CI) P
(N ⫽ 42) (N ⫽ 256)
Child
Breastfeeding 22 (52) 140 (55) 0.91 (0.47–1.8) 0.78 — — —
Male 20 (48) 138 (54) 0.78 (0.40–1.5) 0.45 — — —
Gestational age 40–42 wk 28 (67) 128 (50) 2.00 (1.01–4.0) 0.05 — — —
Birth weight ⬎ 4 kg 16 (38) 52 (20) 2.41 (1.2–4.8) 0.01 2.24 (1.1–4.6) 0.03 1
Environment
Parental atopy 24 (57) 143 (56) 1.05 (0.55–2.0) 0.88 — — —
Maternal smoking 2 (5) 24 (9) 0.48 (0.11–2.1) 0.33 — — —
Born in April to September 28 (67) 132 (52) 1.88 (0.95–3.7) 0.07 2.17 (1.1–4.4) 0.03 1
Day care or siblings 41 (98) 214 (84) 8.05 (1.1–60.1) 0.02 5.80 (0.76–44.4) 0.09 2
High parental education level 38 (91) 186 (73) 3.58 (1.2–10.4) 0.01 2.79 (0.94–8.3) 0.07 1
Hosmer-Lemeshow 2 — — — — 2.74 0.91 —
ROC AUC — — — — 0.72 (0.64–0.80) — 5
CI indicates confidence interval. The prediction rule was as follows: score ⫽ (2 for day care attendance and/or siblings) ⫹ (1 for high parental education level) ⫹ (1 for birth weight of ⬎4
kg) ⫹ (1 for birth in April to September). All variables were dichotomous (0 or 1), and scores ranged from 0 through 5.
dent predictive variables, that is, day TABLE 2 Performance of Different Thresholds for Prediction Rule for RSV LRTI (N ⫽ 298)
care attendance and/or the presence Threshold True-Positive True-Negative Positive Predictive Negative Predictive
Results Results Value, % Value, %
of siblings (odds ratio [OR]: 5.8), high (N ⫽ 42), % (N ⫽ 256), %
parental education level (OR: 2.8), birth ⱖ3 40 (95) 58 (23) 16.8 96.7
weight of ⬎4 kg (OR: 2.2), and month of ⱖ4 33 (79) 148 (58) 23.4 94.3
birth between April and September 5 8 (19) 239 (93) 32.0 87.5
(OR: 2.2) (Table 1). The goodness-of-fit The prediction rule was as follows: score ⫽ (2 for day care attendance and/or siblings) ⫹ (1 for high parental education
level) ⫹ (1 for birth weight of ⬎4 kg) ⫹ (1 for birth in April to September). All variables were dichotomous (0 or 1), and
test indicated an acceptable fit of the scores ranged from 0 through 5.
final prognostic model (P ⫽ .91), and
the AUC was 0.72 (95% confidence in-
terval: 0.64 – 0.80). The shrunk AUC of children in the cohort with and with- tors (8% of all children) had an abso-
was 0.70 (shrinkage factor: 0.97). The out RSV LRTI across different catego- lute risk of 32% (risk ratio: 9.6).
sensitivity analyses with the alterna- ries of risk scores. Figure 2 shows that
children with the lowest scores (0 –2 Furthermore, 62% of the children with
tive outcome of GP-attended RSV
points; 20% of all children) had an ab- RSV LRTI experienced wheezing during
infection and with alternative pre-
dictive factors yielded similar solute risk of 3% for developing RSV the first year of life, compared with
prognostic models with identical dis- LRTI, whereas children with all risk fac- 36% of the children without RSV LRTI
criminating abilities (ROC AUC values (risk ratio: 1.72; P ⫽ .003) (Fig 3). Ex-
of 0.72 and 0.71, respectively). clusion of the episodes that defined
the RSV LRTI group gave similar results
By using the regression coefficients of
the final predictive model, the proba- (59% vs 36%; risk ratio: 1.65; P ⫽ .005).
bility of developing a RSV LRTI can be Children with RSV LRTI used more re-
estimated for each child by using the spiratory drugs at the age of 1 year,
formula given in Table 1. For example, although this finding was not signifi-
a child who is born in July (1 point), cant (15% vs 8%), and more often vis-
attends day care (2 points), has a birth ited a physician because of respiratory
weight of 4.2 kg (1 point), and has par- problems, compared with children
ents who are not highly educated (0 FIGURE 2 without RSV LRTI (48% vs 30%; P ⫽ .03).
Absolute risk to develop RSV LRTI for children
points) has a total score of 1 ⫹ 2 ⫹ with different prediction rule scores. Scores of The HRQoL was lower for children with
1 ⫹ 0 ⫽ 4 points, which corresponds 0, 1, or 2 points (pooled), n ⫽ 60; score of 3 RSV LRTI with respect to 5 of the 10
points, n ⫽ 97; score of 4 points, n ⫽ 116; score
to a probability of developing a RSV of 5 points, n ⫽ 25 were compared by using 2 domains (lungs, stomach, appetite,
LRTI of 23%. Table 2 shows the number test, P ⬍ .001. anxiety, and problem behavior), com-
38 HOUBEN et alARTICLES
To our knowledge, this is the first study
that attempts to predict the risk of
nonhospitalized RSV LRTI for healthy
newborns by using molecular detec-
tion of RSV. Some of our findings de-
serve additional discussion. First, only
341 of the 1080 eligible newborns par-
ticipated in our study, which might
have resulted in selection bias. Com-
parison of the baseline clinical and de-
mographic characteristics between
participants and nonparticipants,
however, showed no differences.
Therefore, we think that our results
are generalizable to all healthy new-
borns. Second, because of the design
of our study, elective cesarean deliver-
ies were overrepresented in this co-
hort (16% vs 6% in the Netherlands).38
Mode of delivery, however, was not as-
sociated with RSV LRTI. Therefore, we
FIGURE 3
Proportions of children with wheezing during the first year of life, for children with and without RSV assume that the results are generaliz-
LRTI in the first year of life. A, Proportion of children with wheezing in each calendar month. Fisher’s able to other modes of delivery. Third,
exact test: November, P ⫽ .001; December, P ⫽ .01. Exclusion of the episodes that defined the RSV LRTI
group yielded similar results (November, P ⫽ .01). B, Cumulative proportion of children with wheezing
the possibility of misclassification at-
in each month of life. Fisher’s exact test or 2 test: all P ⬍ .01, except for month 1 (not significant) and tributable to parental noncompliance
month 5 (P ⫽ .02). Exclusion of the episodes that defined the RSV LRTI group yielded similar results; with recording of respiratory symp-
all P ⬍ .01, except for month 1 (not significant), month 5 (P ⫽ .05), month 6 (P ⫽ .05), and month
7 (P ⫽ .01). toms and collection of nose/throat
swabs cannot be completely ruled out.
However, associations between paren-
pared with children without RSV LRTI strongest predictor in our study (day tal compliance and any potential risk
(Supplemental Figure 4). care attendance and/or the presence factor seem unlikely. Fourth, because
of siblings) is in agreement with the missing values usually do not occur at
DISCUSSION findings of other studies.21–23 High random, exclusion of participants with
We developed a simple prediction rule birth weight may be associated with missing values (complete case analy-
that identifies healthy newborns at delayed parturition and an altered sis) might have resulted in biased es-
high risk of RSV LRTI in the first year of immunologic phenotype.21,34,35 Birth timates.39,40 Therefore, we used impu-
life. Independent predictors for RSV within 6 months before the start of the tation to address the missing values,
LRTI were day care attendance and/or RSV season is a longer window than including missing values for the out-
the presence of siblings, high parental usually found.16,21,23 However, it is come, which yielded results similar to
education levels, birth weight of ⬎4 consistent with the median age of 6 those of the presented complete case
kg, and month of birth between April months for RSV LRTI in the commu- analysis. Fifth, for a number of vari-
and September. nity and/or at GPs in our cohort study ables, we used arbitrary cutoff values
Our prognostic study differs from oth- and in studies by others.1 Highly ed- and/or definitions, mostly in favor of a
ers with respect to the domain and ucated parents might be more care- simple prediction rule or as a result of
outcome studied.21,22 We focused on ful or might seek earlier medical ad- study design. Accessory analyses with
nonhospitalized RSV LRTI among vice if their child develops a alternative cutoff values yielded a sim-
healthy term infants, whereas others respiratory infection.36,37 However, ilar prediction model. Similarly, use of
studied hospitalized RSV in premature parental education levels also may continuous variables (eg, for duration
infants. This may explain the small dif- be associated with other environ- of breastfeeding and number of ciga-
ferences in predictive factors. The mental factors. rettes smoked per day) did not change
PEDIATRICS Volume 127, Number 1, January 2011 39the final model. Therefore, we think als when future RSV vaccines become Houben), the Wilhelmina Children’s
that our prediction rule is robust. available for healthy term infants.44 Hospital Research Fund (grant
The clinical implications of our finding 2004.02), the Catharijne Stichting, and
CONCLUSIONS
include the use of the prediction rule the Dutch Asthma Foundation (grant
by primary care pediatricians, who The risk of RSV LRTI was 10 times higher 3.2.07.001). The funders had no role in
care for the majority of children at risk for children who attended day care, had study design, data collection or analy-
older siblings, had high parental educa- sis, the decision to publish, or prepa-
of and/or with RSV LRTI.1 The incidence
tional levels, had birth weights of ⬎4 kg, ration of the manuscript.
of medically attended RSV infections
and were born between April and Sep-
among children younger than 1 year is We acknowledge Eltje Bloemen, re-
tember, compared with children without
extremely high (⬃44%), and the inci- search nurse, for her participation in
these factors. Clinicians can use these
dence is highest for the group 6 to 12 data collection; Jojanneke Dekkers,
features to differentiate between chil-
months of age (24%).1 Children classi- laboratory technician, for technical as-
dren with high and low risks of RSV LRTI
fied as being at high risk could be mon- sistance with real-time PCR assays;
and subsequently can target preventive
itored more closely and lifestyle Projka Piravalieva-Nikolova, labora-
and monitoring strategies to children at
changes that reduce exposure could tory assistant, and Arthur Gottenkieny,
high risk.
be applied.6–9 When novel preventive laboratory technician, for technical
treatment options become available, ACKNOWLEDGMENTS support; and Hilda Kessel, gynecolo-
these could be used for targeted high- This study was funded by a fellowship gist, and Wouter de Waal, pediatrician,
risk populations.41– 43 Finally, the model award from the European Society for for their assistance with recruitment
may be used in randomized clinical tri- Paediatric Infectious Diseases (to Dr of participants.
REFERENCES
1. Hall CB, Weinberg GA, Iwane MK, et al. The virus: antiviral therapies. Pediatr Infect Dis Palivizumab prophylaxis reduces hospital-
burden of respiratory syncytial virus infec- J. 2003;22(2 suppl):S89 –S92 ization due to respiratory syncytial virus in
tion in young children. N Engl J Med. 2009; 9. van de Pol AC, Rossen JW, Wolfs TF, et al. young children with hemodynamically sig-
360(6):588 –598 Transmission of respiratory syncytial vi- nificant congenital heart disease. J Pediatr.
2. Bourgeois FT, Valim C, McAdam AJ, Mandl KD. rus at the paediatric intensive-care unit: 2003;143(4):532–540
Relative impact of influenza and respiratory a prospective study using real-time PCR. 16. Hall CB. Respiratory syncytial virus and
syncytial virus in young children. Pediatrics. Clin Microbiol Infect. 2010;16(5):488 – 490 parainfluenza virus. N Engl J Med. 2001;
2009;124(6). Available at: www.pediatrics.org/ 10. IMpact-RSV Study Group. Palivizumab, a hu- 344(25):1917–1928
cgi/content/full/124/6/e1072 manized respiratory syncytial virus mono- 17. Holberg CJ, Wright AL, Martinez FD, Ray CG,
3. Stein RT, Sherrill D, Morgan WJ, et al. Respi- clonal antibody, reduces hospitalization Taussig LM, Lebowitz MD. Risk factors for
ratory syncytial virus in early life and risk of from respiratory syncytial virus infection in respiratory syncytial virus-associated
wheeze and allergy by age 13 years. Lancet. high-risk infants. Pediatrics. 1998;102(3): lower respiratory illnesses in the first year
1999;354(9178):541–545 531–537 of life. Am J Epidemiol. 1991;133(11):
1135–1151
4. Bont L, Steijn M, Van Aalderen WM, et al. 11. Ramilo O. Evolution of prophylaxis: MoAb,
Seasonality of long term wheezing follow- siRNA, vaccine, and small molecules. Paedi- 18. Nachman SA, Navaie-Waliser M, Qureshi MZ.
ing respiratory syncytial virus lower respi- atr Respir Rev. 2009;10(suppl 1):23–25 Rehospitalization with respiratory syncytial
virus after neonatal intensive care unit
ratory tract infection. Thorax. 2004;59(6): 12. Bloemers BLP, van Furth AM, Weijerman ME,
discharge: a 3-year follow-up. Pediatrics.
512–516 et al. Down syndrome: a novel risk factor for 1997;100(6). Available at: www.pediatrics.
5. Bont L, Steijn M, Van Aalderen WM, Kimpen respiratory syncytial virus bronchiolitis: a org/cgi/content/full/100/6/e8
JL. Impact of wheezing after respiratory prospective birth-cohort study. Pediatrics.
19. Shay DK, Holman RC, Newman RD, Liu LL,
syncytial virus infection on health-related 2007;120(4). Available at: www.pediatrics.
Stout JW, Anderson LJ. Bronchiolitis-
quality of life. Pediatr Infect Dis J. 2004; org/cgi/content/full/120/4/e1076
associated hospitalizations among US chil-
23(5):414 – 417 13. Simoes EA. Environmental and demo- dren, 1980 –1996. JAMA. 1999;282(15):
6. Contreras PA, Sami IR, Darnell ME, Ottolini graphic risk factors for respiratory syncy- 1440 –1446
MG, Prince GA. Inactivation of respiratory tial virus lower respiratory tract disease. J 20. Stevens TP, Sinkin RA, Hall CB, Maniscalco
syncytial virus by generic hand dishwash- Pediatr. 2003;143(5 suppl):S118 –S126 WM, McConnochie KM. Respiratory syncytial
ing detergents and antibacterial hand 14. Doering G, Gusenleitner W, Belohradsky BH, virus and premature infants born at 32
soaps. Infect Control Hosp Epidemiol. 1999; Burdach S, Resch B, Liese JG. The risk of weeks’ gestation or earlier: hospitalization
20(1):57–58 respiratory syncytial virus-related hospital- and economic implications of prophylaxis.
7. Forbes M. Strategies for preventing respi- izations in preterm infants of 29 to 35 Arch Pediatr Adolesc Med. 2000;154(1):
ratory syncytial virus. Am J Health Syst weeks’ gestational age. Pediatr Infect Dis J. 55– 61
Pharm. 2008;65(23 suppl 8):S13–S19 2006;25(12):1188 –1190 21. Simões EA, Carbonell-Estrany X, Fullarton
8. Jafri HS. Treatment of respiratory syncytial 15. Feltes TF, Cabalka AK, Meissner HC, et al. JR, et al. A predictive model for respiratory
40 HOUBEN et alARTICLES
syncytial virus (RSV) hospitalisation of pre- of childhood respiratory symptom ques- determinants of drug prescription in pri-
mature infants born at 33–35 weeks of ges- tions. Eur Respir J. 1992;5(8):930 –935 mary care. Pharmacoepidemiol Drug Saf.
tational age, based on data from the Span- 29. Fekkes M, Theunissen NC, Brugman E, et al. 2009;18(7):610 – 618
ish FLIP study. Respir Res. 2008;9(1):78 Development and psychometric evaluation 38. Zwart JJ, Richters JM, Ory F, de Vries JI,
22. Rietveld E, Vergouwe Y, Steyerberg EW, of the TAPQOL: a health-related quality of life Bloemenkamp KW, van Roosmalen J. Severe
Huysman MW, de Groot R, Moll HA. Hospital- instrument for 1–5-year-old children. Qual maternal morbidity during pregnancy, de-
ization for respiratory syncytial virus infec- Life Res. 2000;9(8):961–972 livery and puerperium in the Netherlands: a
tion in young children: development of a 30. Theunissen NC, Veen S, Fekkes M, et al. Quality of nationwide population-based study of
clinical prediction rule. Pediatr Infect Dis J. life in preschool children born preterm. Dev Med 371,000 pregnancies. BJOG. 2008;115(7):
2006;25(3):201–207 Child Neurol. 2001;43(7):460–465 842– 850
23. Law BJ, Langley JM, Allen U, et al. The Pedi- 31. Royston P, Moons KG, Altman DG, Vergouwe 39. Sterne JA, White IR, Carlin JB, et al. Multiple
atric Investigators Collaborative Network Y. Prognosis and prognostic research: de- imputation for missing data in epidemiolog-
on Infections in Canada study of predictors veloping a prognostic model. BMJ. 2009;338: ical and clinical research: potential and pit-
of hospitalization for respiratory syncytial b604 falls. BMJ. 2009;338:b2393
virus infection for infants born at 33 32. Altman DG, Vergouwe Y, Royston P, Moons 40. Greenland S, Finkle WD. A critical look at
through 35 completed weeks of gestation. KG. Prognosis and prognostic research: val- methods for handling missing covariates in
Pediatr Infect Dis J. 2004;23(9):806 – 814 idating a prognostic model. BMJ. 2009;338:
epidemiologic regression analyses. Am J
b605
24. Landon MB, Spong CY, Thom E, et al. A mul- Epidemiol. 1995;142(12):1255–1264
ticenter, randomized trial of treatment for 33. Harrell FE Jr, Lee KL, Mark DB. Multivariable
41. Olszewska W, Openshaw P. Emerging drugs
mild gestational diabetes. N Engl J Med. prognostic models: issues in developing
for respiratory syncytial virus infection. Ex-
2009;361(14):1339 –1348 models, evaluating assumptions and ade-
pert Opin Emerg Drugs. 2009;14(2):207–217
quacy, and measuring and reducing errors.
25. Ochola R, Sande C, Fegan G, et al. The level and Stat Med. 1996;15(4):361–387 42. DeVincenzo J, Cehelsky JE, Alvarez R, et al.
duration of RSV-specific maternal IgG in infants Evaluation of the safety, tolerability and
34. Atègbo JM, Grissa O, Yessoufou A, et al. Mod-
in Kilifi, Kenya. PLoS One. 2009;4(12):e8088 pharmacokinetics of ALN-RSV01, a novel
ulation of adipokines and cytokines in ges-
26. Ermers MJ, Rovers MM, van Woensel JB, tational diabetes and macrosomia. J Clin RNAi antiviral therapeutic directed against
Kimpen JL, Bont LJ. The effect of high dose Endocrinol Metab. 2006;91(10):4137– 4143 respiratory syncytial virus (RSV). Antiviral
inhaled corticosteroids on wheeze in in- 35. Moore TR. A comparison of amniotic fluid Res. 2008;77(3):225–231
fants after respiratory syncytial virus fetal pulmonary phospholipids in normal 43. DeVincenzo J, Lambkin-Williams R, Wilkin-
infection: randomised double blind placebo and diabetic pregnancy. Am J Obstet Gy- son T, et al. A randomized, double-blind,
controlled trial. BMJ. 2009;338:b897 necol. 2002;186(4):641– 650 placebo-controlled study of an RNAi-based
27. van der Zalm MM, van Ewijk BE, Wilbrink B, 36. Saunders NR, Tennis O, Jacobson S, Gans M, therapy directed against respiratory syncy-
Uiterwaal CS, Wolfs TF, van der Ent CK. Re- Dick PT. Parents’ responses to symptoms of tial virus. Proc Natl Acad Sci U S A. 2010;
spiratory pathogens in children with and respiratory tract infection in their children. 107(19):8800 – 8805
without respiratory symptoms. J Pediatr. CMAJ. 2003;168(1):25–30 44. Empey KM, Peebles RS Jr, Kolls JK. Pharma-
2009;154(3):396 – 400 37. de Jong BM, van der Ent CK, van der Zalm cologic advances in the treatment and pre-
28. Brunekreef B, Groot B, Rijcken B, Hoek G, MM, et al. Respiratory symptoms in young vention of respiratory syncytial virus. Clin
Steenbekkers A, de Boer A. Reproducibility infancy: child, parent and physician related Infect Dis. 2010;50(9):1258 –1267
PEDIATRICS Volume 127, Number 1, January 2011 41You can also read
