Is preterm birth associated with the development of wheezing disorders during childhood?
Two reviewers will independently search PubMed, EMBASE, TRIP database, and Google Scholar for eligible studies. We will identify additional studies by screening reference lists and citations of articles of interest via ISI Web of Knowledge. We will also approach an international panel of experts to identify ongoing and unpublished studies, as well as studies that might have been missed by our search strategy.
Types of study to be included
We will consider the following observational designs for inclusion: cohort, case-control, and cross-sectional studies. We do not expect to identify experimental or quasi-experimental studies relevant to the research question. However if they are identified we will consider their inclusion if relevant. Two reviewers will independently assess study eligibility. The final decision to include will be based on consensus and arbitration by a third author if necessary. When studies have significantly overlapping populations, one study will be selected for final inclusion, study size being taken into account as a major determinant.
Condition or domain being studied
Wheezing disorders during childhood, defined as a diagnosis or reported symptoms of wheezing or asthma.
Children aged 0.5-18 years. Only studies where at least 50% of the cohort is born from 1995 onwards, and none of the children is born before 1990, are eligible. The rationale for this is that important practice changes in neonatal medicine have contributed to improved outcomes among (very) preterm infants in the last two decades. To obtain an estimate that is relevant to current neonatal practice, we have set this pragmatic time constraint.
The exposure under study is preterm birth, defined as being born at a gestational age less than 37 completed weeks. Studies comparing the outcome between one or more subgroups at the more extreme end of gestation (e.g. <32 weeks) and children born at term are also eligible, as long as the comparison group inclusion criteria are met.
The comparison group will consist of children born at term, defined as a gestational age of 37 weeks or above (>=37 weeks). Studies using broader gestational age definitions to define term births (e.g. >= 36 weeks) will not be considered. Post-term births (>42 weeks) will be excluded when possible.
The primary outcome is any wheezing disorder.
We will use the following hierarchy to select the outcome from studies reporting multiple outcomes: asthma; persistent wheezing; recurrent wheezing; severe wheezing; and wheezing.
We will use the following hierarchy to select the outcome with the highest level of ascertainment: clinician diagnosis; documented medication use as a wheezing disorders proxy; routinely collected health care data; parental or patient-reported physician diagnosis; parental or patient-reported medication use; and parental or patient-reported symptoms.
‘Ever’ wheezing/asthma will be favoured over recent and current wheezing disorders, respectively.
Data extraction, (selection and coding)
Two reviewers will independently extract the following data from selected studies:authors; full reference; study design; study location; sample size; inclusion- and exclusion criteria; age range and birth year of study participants; method of ascertainment of exposure and outcomes; outcome measure; unadjusted and adjusted association measures. Unadjusted association measures will preferably be calculated using crude data via production of 2x2 tables.
If available, adjusted association estimates of a linear relationship between gestational age and wheezing disorder risk will be extracted to assess a possible “dose-response relationship” between degree of prematurity and risk of wheezing disorders.
Risk of bias (quality) assessment
Two reviewers will independently assess risk of bias for each included study using the Effective Public Health Practice Project quality assessment tool for quantitative studies. For this purpose the following covariates are considered to be the main confounders of the association between preterm birth and wheezing disorders: gender; maternal smoking during pregnancy; maternal atopy or asthma, or a family history of atopy or asthma. Each study will be assigned to one of the following categories: low risk of bias, moderate risk of bias, and high risk of bias. Again, final judgment of risk of bias will be based on consensus with arbitration by a third author if necessary.
Strategy for data synthesis
Unadjusted association measures will be pooled via Mantel-Haenszel analysis. We anticipate important heterogeneity among the studies and will perform all analyses using a random effects model. Adjusted effect estimates will be pooled using generic inverse variance. Standard errors for the individual study point estimates will be calculated from the respective confidence intervals as described in the Cochrane Handbook (chapter 7). When studies report adjusted association measures for several gestational age strata, the least preterm stratum will be selected for the comparison so as to arrive at the most conservative estimate. Adjusted effect estimates of a dose-response relationship between gestational age and wheezing disorders will be calculated for individual studies that reported multiple gestational age strata using a fixed-effects log-linear dose-response regression model. Individual study summary estimates will then be pooled using generic inverse variance. Between-study heterogeneity will be assessed using Q-statistic and the I-squared test. We will perform meta-regression analysis using the following study characteristics to assess their contribution to overall study heterogeneity: study size, population age, diagnosis ascertainment, wheezing type, and publication year. Small study effects will be assessed using funnel plots, and tested using Harbord’s test for unadjusted association measures, and Egger’s test for adjusted association measures. If possible, population attributable risks will be calculated using the data as described by Rockhill et al (Am J Public Health 1998). All analyses will be performed using Stata 12 software.
Analysis of subgroups or subsets
Subgroup analyses will be performed based on age (<4 years vs. >=4 years), and degree of prematurity (<32 weeks vs. 32-26 weeks). We will also perform sensitivity analyses according to study size (n<10,000 vs. n>=10,000), risk of bias (low, moderate, or high), and outcome definition (asthma vs. wheezing) and ascertainment (clinician diagnosis and/or medication use vs. parental-reported outcomes).
We will present our findings in a single manuscript to be submitted for publication to an international peer-reviewed medical journal.
Contact details for further information
Edinburgh EH8 9AG
Organisational affiliation of the review
School for Public Health and Primary Care (CAPHRI), Maastricht University
Dr Jasper Been, School for Public Health and Primary Care (CAPHRI), Maastricht University, Maastricht, Netherlands Miss Marlies Lugtenberg, School for Public Health and Primary Care (CAPHRI), Maastricht University, Maastricht, Netherlands Miss Eline Smets, School for Public Health and Primary Care (CAPHRI), Maastricht University, Maastricht, Netherlands Professor Constant van Schayck, School for Public Health and Primary Care (CAPHRI), Maastricht University, Maastricht, Netherlands Professor Boris Kramer, Department of Paediatrics, Maastricht University Medical Centre, Maastricht, Netherlands Dr Monique Mommers, School for Public Health and Primary Care (CAPHRI), Maastricht University, Maastricht, Netherlands Professor Aziz Sheikh, Centre for Population Health Sciences, The University of Edinburgh, Edinburgh, United Kingdom
Anticipated or actual start date
01 April 2013
Anticipated completion date
01 August 2013
JB is supported by a Maastricht University Kootstra Talent Fellowship
JB is supported by the International Pediatric Research Foundation Young Investigators Exchange Programme
Formal screening of search results against eligibility criteria
Risk of bias (quality) assessment
PROSPERO This information has been provided by the named contact for this review. CRD has accepted this information in good faith and registered the review in PROSPERO. CRD bears no responsibility or liability for the content of this registration record, any associated files or external websites.