Preterm birth and childhood wheezing disorders: a systematic review and meta-analysis

Abstract

Background: Accumulating evidence implicates early life factors in the aetiology of non-communicable diseases, including asthma/wheezing disorders. We undertook a systematic review investigating risks of asthma/wheezing disorders in children born preterm, including the increasing numbers who, as a result of advances in neonatal care, now survive very preterm birth.

Methods and findings: Two reviewers independently searched seven online databases for contemporaneous (1 January 1995-23 September 2013) epidemiological studies investigating the association between preterm birth and asthma/wheezing disorders. Additional studies were identified through reference and citation searches, and contacting international experts. Quality appraisal was undertaken using the Effective Public Health Practice Project instrument. We pooled unadjusted and adjusted effect estimates using random-effects meta-analysis, investigated "dose-response" associations, and undertook subgroup, sensitivity, and meta-regression analyses to assess the robustness of associations. We identified 42 eligible studies from six continents. Twelve were excluded for population overlap, leaving 30 unique studies involving 1,543,639 children. Preterm birth was associated with an increased risk of wheezing disorders in unadjusted (13.7% versus 8.3%; odds ratio [OR] 1.71, 95% CI 1.57-1.87; 26 studies including 1,500,916 children) and adjusted analyses (OR 1.46, 95% CI 1.29-1.65; 17 studies including 874,710 children). The risk was particularly high among children born very preterm (<32 wk gestation; unadjusted: OR 3.00, 95% CI 2.61-3.44; adjusted: OR 2.81, 95% CI 2.55-3.12). Findings were most pronounced for studies with low risk of bias and were consistent across sensitivity analyses. The estimated population-attributable risk of preterm birth for childhood wheezing disorders was ≥3.1%. Key limitations related to the paucity of data from low- and middle-income countries, and risk of residual confounding.

Conclusions: There is compelling evidence that preterm birth-particularly very preterm birth-increases the risk of asthma. Given the projected global increases in children surviving preterm births, research now needs to focus on understanding underlying mechanisms, and then to translate these insights into the development of preventive interventions.

Review registration: PROSPERO CRD42013004965.

Figure 1. Flowchart outlining study selection.

WHO GHL, World Health Organization Global Health Library; WHOLIS, World Health Organization Library Information System.

Figure 2. Meta-analysis of unadjusted association between preterm birth and childhood wheezing disorders.

Heterogeneity: I ² = 82% (95% CI 75%–87%). PTB, preterm birth; WD, wheezing disorders.

Figure 3. Meta-analysis of adjusted association between preterm birth and childhood wheezing disorders.

Subgroups taken from individual studies noted in parentheses. Heterogeneity: I ² = 80% (95% CI 68%–86%). Individual study adjustment for the primary confounders is depicted. Additional confounders adjusted for are outlined in Table S3. FH, family history; LRI, lower respiratory infection; PTB, preterm birth; WD, wheezing disorders.

Figure 4. Meta-analysis of adjusted dose–response association between gestational age (per week increase) and childhood wheezing disorders.

Heterogeneity: I ² = 90% (95% CI 85%–92%). Individual study adjustment for the primary confounders is depicted. Additional confounders adjusted for are outlined in Table S3. FH, family history; WD, wheezing disorders.

Figure 5. Meta-analysis of association between very preterm birth and childhood wheezing disorders.

(A) unadjusted effect estimates; (B) adjusted effect estimates. Subgroups taken from individual studies noted in parentheses. Heterogeneity: I ² (unadjusted) = 62% (95% CI 9%–79%); I ² (adjusted) = 0% (95% CI 0%–68%). Confounders adjusted for in individual studies are outlined in Figure 3 and Table S3. LRI, lower respiratory infection; PTB, preterm birth; WD, wheezing disorders.

Figure 6. Meta-analysis of association between moderately preterm birth and childhood wheezing disorders.

(A) unadjusted effect estimates; (B) adjusted effect estimates. Subgroups taken from individual studies noted in parentheses. Heterogeneity: I ² (unadjusted) = 92% (95% CI 87%–94%); I ² (adjusted) = 89% (95% CI 82%–93%). Confounders adjusted for in individual studies are outlined in Figure 3 and Table S3. LRI, lower respiratory infection; PTB, preterm birth; WD, wheezing disorders.

Figure 7. Funnel plots for studies reporting unadjusted and adjusted association measures.

(A) unadjusted association measures; (B) adjusted association measures.