- Split View
-
Views
-
Cite
Cite
David P Burgner, Jessica E Miller, Tobias Strunk, Natasha Nassar, Growth Parameters, Effect Measure Modification and the Association Between Vaccination and Early Childhood Hospitalization With Non-targeted Infections, Clinical Infectious Diseases, Volume 66, Issue 2, 15 January 2018, Pages 318–319, https://doi.org/10.1093/cid/cix838
- Share Icon Share
To the Editor—We read with interest the recent findings of Bardenheier et al [1] that report a reduced risk of hospitalization for non-targeted infections (NTI) among children aged 16 to 24 months. As the authors note, some of the unadjusted and adjusted hazard ratios of NTI hospitalization risk for the last type of vaccine received were in opposite directions, an observation that warrants further investigation. Compared to inactivated vaccines, if the last vaccine type received was either a live vaccine, or concurrent live and inactivated vaccines, then NTI risk was increased in unadjusted analysis but was decreased in adjusted models. These results suggest that effect measure modification or confounding by 1 or more factors may have important implications for the associations and interpretation of the findings.
In particular, oversimplification of the variable summarizing prematurity and/or low birth weight may have resulted in possible effect measure modification being overlooked. The authors used the child’s ICD-9 code 765.x (disorders relating to short gestation and low birth weight) and the maternal ICD-9 codes 644.20 and 644.21 (early onset of delivery, < 37 completed weeks gestation) to dichotomize for prematurity and/or low birth weight (< 2500 grams). A more nuanced approach is warranted if more granular data are available.
Reduced gestational age and suboptimal intrauterine growth are important potential confounders and/or mediators of the association between vaccine types and NTI hospitalizations. Age at vaccination, immunological response to vaccination, and risk of hospitalization may all differ by gestational age and birth weight [2–4].
Dichotomization of prematurity/low birth weight may obscure important clinically relevant effects that modify the relationship between vaccine type and NTI hospitalization. In our total population study of ~720000 singleton births in Western Australia, we found that even late preterm and early term birth, and near normal birth weight have marked and pervasive associations with risk of infection-related hospitalization throughout childhood. Hospitalization rates increased by 12% for each reduction in week of gestation < 39–40 weeks (rate ratio 1.12, 95% confidence interval 1.12–1.13), by 19% for each 500-g reduction in birth weight < 3000–3500 grams (1.19, 1.18–1.21), and by 41% for each 5-cm reduction in birth length < 45–50 cm (1.41, 1.38–1.45) [5].
Therefore, further analysis with stratification for gestational age and growth parameters at birth may provide insight into the potential mechanisms and highlight subgroups who may benefit most. In low- and middle-income populations, in whom the suggested nonspecific beneficial effects of vaccines may have the greatest impact, moderate prematurity and suboptimal intrauterine growth are common. More detailed analysis of the effects of these parameters on the relationships between vaccine type and NTI is important to increase validity and rigor of the findings and inform translation.
Notes
Acknowledgments. D. P. B. is supported by a National Health and Medical Research Council (Australia) Senior Research Fellowship (APP1064629) and is a National Heart Foundation (Australia) Honorary Future Leader Fellow (100026)
Potential conflicts of interest. All authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
Comments