Journal reference: Eick AA, Uyeki TM, Klimov A, et al. Maternal influenza vaccination and effect on influenza virus infection in young infants. Arch Pediatr Adolesc Med. 2010;0(2010):archpediatrics.2010.192 ^[1]

Link: http://dx.doi.org/10.1001/archpediatrics.2010.192

Published: 4 October 2010

Evidence cookie says...

Maternal influenza vaccination during pregnancy has a limited effect on infant influenza-like illness (ILI). it was not associated with a change in the incidence of medically attended ILI in infants up to 6-months of age; however, it was associated with a ↓ 39% risk of hospital admissions for ILI Australian guidelines already recommend influenza vaccination in pregnancy [2]. Note: this study has significant limitations, please see below for details.

More details:

Article details

Study design:

non-randomised prospective cohort study

Study aim:

to assess the effect of influenza vaccination during pregnancy on laboratory-confirmed influenza in infants to 6 months of age

Methods summary:

• Conducted on the Navajo and White Mountain Apache Indian reservations
• Conducted between 2002- 2005 over three influenza seasons
• Inclusion was restricted to mothers who delivered a healthy infant at 36 week’s or later gestation and infants who were aged 2 weeks or younger
• Recruitment of mother-infant pairs were after delivery at the hospital or by home visit; for all subjects the study started after delivery (decision for influenza vaccination was made by the treating clinician and pregnant woman)
• Primary outcome measures:
  • Medically attended influenza-like illness (ILI)
    • at least one of the of the following: fever of 38.0 °C or higher, diarrhoea, or respiratory symptoms
• Laboratory confirmation of influenza performed by nasopharyngeal aspirate with rapid diagnostic testing, isolation of influenza virus or serology

Results summary:

• ILI incidence rate was 7.2 per 1000 person days for infants of unvaccinated mothers (note comment in methodological weaknesses)
• ILI incidence rate was 6.7 per 1000 person days for infants of vaccinated mothers (note comment in methodological weaknesses)
• 41% reduction in risk of laboratory confirmed influenza virus infection for infants of vaccinated mothers
  • relative risk [RR] = 0.59, 95% confidence interval [CI] 0.37-0.93
• Maternal influenza vaccination did not confer a statistically significant effect on ILI
  • RR = 0.92, 95% CI 0.73-1.16
• However, there was a 39% reduction in risk of ILI hospitalisation for infants of vaccinated mothers
  • RR = 0.58, 95% CI 0.41-0.83
• Infants of vaccinated mothers had higher antibody titres at birth and 2-3 months for all 8 strains investigated

Study conclusion:

Maternal influenza vaccination was significantly associated with reduced risk of influenza virus infection and hospitalisation for an ILI up to 6 months of age, and increased influenza antibody titres in infants through 2 to 3 months of age.

Participants:

• Conducted on the Navajo and White Mountain Apache Indian reservations in the Southwest region of the United States of America
  • Health care administered through the Indian Health Service
  • 6 Navajo hospitals and 1 White Mountain Apache reservation hospital participated in the study
• 1160 mother-infant pairs from Indian reservations
• 49% of infants were born to mothers who received influenza vaccine during their pregnancy
• The influenza-vaccinated mothers were more likely to have breastfed their infants compared to the unvaccinated mothers (81% vs 74%, P = 0.007)

Methodological weaknesses

• Maternal vaccination status based upon maternal recall if not documented in medical records
• Narrow spectrum of patients, limiting applicability to the general population
• Less than 10% of eligible pregnant women received influenza vaccine prior to the study, but during the study 49% of mothers were vaccinated which may reflect effect of education campaign prior to the study.
• Study did not evaluate maternal influenza exposures or illness during pregnancy or after delivery
• Limitations of sensitivity of serological diagnosis of influenza
• Recruitment procedure is not clearly described; one wonders about potential recruitment bias.
• There was a statistically (and arguably, clinically) significant difference between vaccinated and unvaccinated mothers in the rate of breastfeeding (81% vs 74%, P = 0.007).
  • Although potential confounders including breastfeeding were planned to be “investigated”, the univariate regression model was not adjusted for breastfeeding.  The authors claim that they found no statistical significant associations between breastfeeding and “the occurrence of any of the outcomes” but no data is given to support this assertion.
• Wide confidence interval on the relative risk reduction of laboratory confirmed influenza
  • the study is relatively underpowered to examine this
• The unqualified reporting of the raw ILI incidence rates of infants born to unvaccinated and vaccinated women (7.2 and 6.7 per 1000 person-days) in the article abstract is misleading.
  • It suggests that the study detected a lower rate of ILI in vaccinated women when in fact no statistical difference was detected [RR = 0.92, 95% CI 0.73-1.16]

Methodological strengths

• Laboratory confirmation of infant influenza
• Conducted over three influenza seasons
• Study commenced after delivery and had no impact on decision to vaccinate during pregnancy
• Potential confounders such as smoking, day care attendance, wood /coal stoves and infant breastfeeding were considered

Biases and conflicts of interests

• O’Brien KL (final author) received research project grants unrelated to this article from MedImmune and Pfizer and serves on expert advisory committees for pneumococcal vaccines from Sanofi-Pasteur.  All three companies manufacture influenza vaccine.
• Funding was received from Aventis-Pasteur and Evans-Powderject (manufacturer of vaccines)

Clinical relevance to primary health care

The conclusions of this study need to be treated cautiously.  The method of participant selection and recruitment raises the distinct possibility of introduction of bias and uncontrolled confounding factors.  The study participants were all from Navajo and White Mountain Apache Indian reservations.  There are significant potential funding biases and conflicts of interests.  Moreover, the authors overstate the clinical significance of their findings, or at least, it is inferred.

The primary outcome of the study is medically attended influenza-like illness (ILI) in infants up to 6-months of age.  The study does not demonstrate a statistical difference in the incidence of the primary outcome between vaccinated and unvaccinated mothers.  This finding of the primary outcome is not clearly reported and is missing from the results section of the abstract.  Misleading raw incidence rates without indication of its statistical significance are given instead.

The study does seem to demonstrate that maternal vaccination was associated with a lower risk of hospitalisation for ILI.  However, the non-randomised and non-blinded nature of the study needs to be considered.

The study also demonstrates that maternal vaccination was associated with a lower risk of laboratory confirmed influenza infection.  However, this is of unclear clinical significance in the context that there was no statistical difference in the incidence of ILI between the two groups.

At a certain pragmatic level, the study findings are clinically moot as influenza vaccination is already recommended for pregnant women who will be in the second or third trimester during influenza season ^[2] because of the risk of influenza complications in pregnancy.  It is also recommended that it be offered in advance to women planning a pregnancy ^[2].

A reasonable “take home message” for Australian general practitioners is that it is possible that maternal influenza vaccination may give limited protection to infants.  This should not be used as justification for maternal influenza vaccination, insofar that there are better clinical reasons to do so.

References

1. Eick AA, Uyeki TM, Klimov A, et al. Maternal influenza vaccination and effect on influenza virus infection in young infants. Arch Pediatr Adolesc Med. 2010;0(2010):archpediatrics.2010.192
2. The Australian Immunisation Handbook 9^th Edition. Australian Government Department of Health and Aging. 2008

Editor: Michael Tam