Re: Maternal vaccination against H1N1 influenza and offspring mortality: population based cohort study and sibling design
We would like to thank Junya et al. for their interest in our paper. In response to their concerns we have re-calculated the risk of offspring mortality so that also offspring to women undergoing vaccination 0-1 months before conception were classified as exposed. This did not change the risk estimates more than marginally (e.g. the HR for fetal death went from 0.83 to 0.86 in the population-based analysis; and from 0.88 to 0.90 in the sibling analysis).
In three regions (Kalmar, Norrbotten and Värmland), exposure classification was not possible in a subset of individuals, who could not be separated from the truly unexposed women. Consequently these women were classified as unexposed. However, we carried out a sensitivity analysis where these individuals were excluded (this did not affect the overall risk estimates). Junya et al. suggest that individuals with uncertain exposure in these three regions instead be classified as ”exposed” in the three regions. To do that we would need to classify all unexposed women in these regions as exposed.
Overall, some 16.2% of all offspring were exposed in our study. In the three regions Kalmar, Norrbotten and Värmland this figure was instead 14.0%. However, since we cannot identify the women who did not consent, and cannot distinguish them from those who did not undergo vaccination, we abstained from the suggested sensitivity analysis. Re-classifying all these women as exposed is instead likely to introduce bias. Of note, women from the three regions Kalmar, Norrbotten and Värmland represented 13% of the total study population.
Finally, we agree with the Junya et al. that preterm birth, low birth weight small for gestational age and spontaneous abortion are important outcomes. We have already published data on H1N1 vaccination and these outcomes from a large regional cohort in Sweden [1] and therefore abstained from these analyses in the current study [2].
1. Ludvigsson JF, Zugna D, Cnattingius S, et al. Influenza H1N1 vaccination and adverse pregnancy outcome. Eur J Epidemiol 2013;28(7):579-88 doi: 10.1007/s10654-013-9813-z[published Online First: Epub Date]|.
2. Ludvigsson JF, Strom P, Lundholm C, et al. Maternal vaccination against H1N1 influenza and offspring mortality: population based cohort study and sibling design. BMJ 2015;351:h5585 doi: 10.1136/bmj.h5585[published Online First: Epub Date]|.
Competing interests:
No competing interests
01 February 2016
Jonas F. Ludvigsson
Professor of clinical epidemiology and paediatrician
Peter Ström, Cecilia Lundholm, Sven Cnattingius, Anders Ekbom, Åke Örtqvist, Nils Feltelius, Fredrik Granath, Olof Stephansson
Karolinska Institutet
Department of Medical Epidemiology and Biostatistics
Rapid Response:
Re: Maternal vaccination against H1N1 influenza and offspring mortality: population based cohort study and sibling design
We would like to thank Junya et al. for their interest in our paper. In response to their concerns we have re-calculated the risk of offspring mortality so that also offspring to women undergoing vaccination 0-1 months before conception were classified as exposed. This did not change the risk estimates more than marginally (e.g. the HR for fetal death went from 0.83 to 0.86 in the population-based analysis; and from 0.88 to 0.90 in the sibling analysis).
In three regions (Kalmar, Norrbotten and Värmland), exposure classification was not possible in a subset of individuals, who could not be separated from the truly unexposed women. Consequently these women were classified as unexposed. However, we carried out a sensitivity analysis where these individuals were excluded (this did not affect the overall risk estimates). Junya et al. suggest that individuals with uncertain exposure in these three regions instead be classified as ”exposed” in the three regions. To do that we would need to classify all unexposed women in these regions as exposed.
Overall, some 16.2% of all offspring were exposed in our study. In the three regions Kalmar, Norrbotten and Värmland this figure was instead 14.0%. However, since we cannot identify the women who did not consent, and cannot distinguish them from those who did not undergo vaccination, we abstained from the suggested sensitivity analysis. Re-classifying all these women as exposed is instead likely to introduce bias. Of note, women from the three regions Kalmar, Norrbotten and Värmland represented 13% of the total study population.
Finally, we agree with the Junya et al. that preterm birth, low birth weight small for gestational age and spontaneous abortion are important outcomes. We have already published data on H1N1 vaccination and these outcomes from a large regional cohort in Sweden [1] and therefore abstained from these analyses in the current study [2].
1. Ludvigsson JF, Zugna D, Cnattingius S, et al. Influenza H1N1 vaccination and adverse pregnancy outcome. Eur J Epidemiol 2013;28(7):579-88 doi: 10.1007/s10654-013-9813-z[published Online First: Epub Date]|.
2. Ludvigsson JF, Strom P, Lundholm C, et al. Maternal vaccination against H1N1 influenza and offspring mortality: population based cohort study and sibling design. BMJ 2015;351:h5585 doi: 10.1136/bmj.h5585[published Online First: Epub Date]|.
Competing interests: No competing interests