Abstract
Background
Influenza vaccination during pregnancy reduces influenza-associated illness in the women and their infants, but effectiveness estimates against influenza-associated hospitalization are limited and lacking from settings with high human immunodeficiency virus (HIV) infection prevalence. We assessed the effect of maternal vaccination in HIV-uninfected women and women with HIV in preventing influenza-associated hospitalizations in infants and the women.
Methods
During 2015–2018, influenza vaccination campaigns targeting pregnant women were augmented at selected antenatal clinics; these were coupled with prospective hospital-based surveillance for acute respiratory or febrile illness in infants aged <6 months and cardiorespiratory illness among pregnant or postpartum women. Vaccine effectiveness (VE) was assessed using a test-negative case-control study.
Results
Overall, 71 influenza-positive and 371 influenza-negative infants were included in the analysis; mothers of 26.8% of influenza-positive infants were vaccinated during pregnancy compared with 35.6% of influenza-negative infants, corresponding to an adjusted VE (aVE) of 29.0% (95% confidence interval [CI], −33.6% to 62.3%). When limited to vaccine-matched strains, aVE was 65.2% (95% CI, 11.7%–86.3%). For maternal hospitalizations, 56 influenza-positive and 345 influenza-negative women were included in the analysis, with 28.6% of influenza-positive women being vaccinated compared with 38.3% of influenza-negatives, for an aVE of 46.9% (95% CI, −2.8% to 72.5%). Analysis restricted to HIV-uninfected women resulted in 82.8% (95% CI, 40.7%–95.0%) aVE. No significant aVE (−32.5% [95% CI, −208.7% to 43.1%]) was detected among women with HIV.
Conclusions
Influenza vaccination during pregnancy prevented influenza-associated hospitalizations among young infants when infected with vaccine strains and among HIV-uninfected women.
Keywords: influenza, pregnancy, vaccine
Influenza vaccination during pregnancy prevented influenza-associated hospitalizations among young infants when infected with influenza vaccine strains, and among women without HIV, although, no effect was detected in women living with HIV.
Randomized controlled trials (RCTs) performed in low- and middle-income countries (LMICs) have demonstrated that the efficacy of seasonal influenza vaccination during pregnancy is 50%–70% in preventing mild to moderately severe laboratory-confirmed influenza-associated illness in pregnant women [1, 2]. After vaccination, maternal antibodies cross the placenta and may offer protection to the infant during the first months of life. Therefore, an additional benefit of maternal vaccination has been the protection of the young infants. Maternal vaccination to protect young infants is pertinent since there is no influenza vaccine approved for use in infants aged <6 months. A meta-analysis of 4 RCTs in pregnant women revealed a combined vaccine efficacy of 36% (95% confidence interval [CI], 22%–48%) against laboratory-confirmed influenza-associated illness in infants aged <6 months [1–5]. Vaccine efficacy was higher if the analyses were restricted to the first 8 weeks (86%) or 4 months (68%) of life [2, 6].
From the previous RCTs, only a South African trial assessed vaccine immunogenicity and efficacy among pregnant women living with human immunodeficiency virus (WLWH) and their infants [1]. In that study, despite WLWH having lower levels of neutralization, hemagglutination inhibition, and H1-hemagglutinin stalk domain antibodies following vaccination compared with those without human immunodeficiency virus (HIV) [7–9], the overall vaccine efficacy against influenza-associated illness was similar in both groups (57.7% and 50.4%, respectively) [1]. The study in WLWH was not powered to detect vaccine efficacy in the HIV-exposed infants; nevertheless, similar influenza attack rates were detected in infants born to WLWH in the vaccinated (5.0%) or placebo (6.8%) groups. Attack rates were lower in infants of HIV-uninfected placebo recipients (3.6%) [1].
The low incidence of influenza-associated hospitalization within each influenza season poses a challenge to demonstrate the effectiveness of influenza vaccination against severe disease. Nonetheless, a few observational studies, mainly from Europe and the United States (US), have reported a 45%–92% reduction of laboratory-confirmed influenza-associated hospitalizations in infants aged <6 months by vaccinating pregnant women [3, 10–15]. Furthermore, a multicountry study conducted in Australia, Canada, Israel, and the US reported a vaccine effectiveness (VE) of 40% (95% CI, 12%–59%) against influenza-associated hospitalizations in pregnant women from 2010 to 2016 [16].
Understanding the effect of influenza vaccination on severe influenza outcomes in infants and during pregnancy in LMICs, including among WLWH and their infants, will help to delineate the potential benefit of the maternal immunization programs in these countries. We conducted a prospective test-negative study over 4 consecutive influenza seasons. The primary objective was to evaluate the effectiveness of influenza vaccination of pregnant women in reducing the risk of polymerase chain reaction (PCR)–confirmed influenza-associated hospitalization in their infants during the first 6 months of life, and to estimate VE in pregnant or postpartum women.
METHODS
From 2015 to 2018, once the inactivated influenza vaccine (IIV) formulated for the Southern Hemisphere was available in South Africa, we provided additional IIV doses to selected antenatal clinics in 2 South African cities (Johannesburg and Cape Town), aiming to increase vaccination coverage of pregnant women attending those clinics to at least 50% [17]. Study staff based at the clinics compiled vaccination registers of the women attending the clinics during the period that IIV was available. Vaccination was also documented at the discretion of the attending nurse on the individual maternal antenatal cards. Prospective active surveillance was conducted among infants aged <6 months hospitalized for acute respiratory or febrile illness (ARI/FI), from when IIV was available to the end of each calendar year, at 4 hospitals in the 2 cities; and among pregnant women and women within 42 days postdelivery hospitalized for cardiorespiratory illness in 6 hospitals during the influenza seasons. More details on study design and population are shown in the Supplementary material. Enrolled participants had respiratory swabs collected and tested by real-time PCR for influenza as previously described [1, 18]. Influenza A–positive samples were subtyped as A(H1N1)pdm09 or A(H3N2). All A(H1N1)pdm09 viruses were considered to be vaccine-matched as they belonged to the 6B lineage with antigenic characteristics similar to vaccine strains; for A(H3N2) each year we compared the dominant lineage circulating in South Africa with the vaccine lineage and considered them matched or unmatched (Supplementary Table 1) [19–23]. Influenza B–positive samples were typed as Victoria or Yamagata lineages and compared with the vaccine lineages (Supplementary Table 1).
Influenza Vaccination
During the study period, only trivalent IIV formulations were available. The virus strains contained in the vaccines recommended for the Southern Hemisphere for each year are shown in Supplementary Table 1. Maternal immunization status was determined based on the information in the antenatal cards, if these were available at the time of admission, or study vaccination registers [17]. Women were considered vaccinated if there was written documentation of receipt of IIV during pregnancy and if they were vaccinated ≥14 days before giving birth (infant analysis) or ≥14 days before hospital admission (maternal analysis). Women vaccinated <14 days before birth/admission and those who had unknown vaccination status or had a missing vaccination date were excluded from analyses.
Statistical Analysis
The primary objective of the study was to estimate VE against influenza-associated hospitalization in infants aged <6 months. Assuming 5% significance level, 80% power, case-control ratio ranging from 1:4 to 1:8, vaccine coverage of 50%, and a VE of 60%, we estimated a minimum sample size of 46 cases.
Two unmatched test-negative case-control analyses for infants and women were used to estimate VE: (1) For infants, cases were infants hospitalized with ARI/FI, whose mothers were eligible to have received IIV during pregnancy, aged <6 months at hospital admission with PCR-confirmed influenza; for each infant case, 5 controls were randomly selected among the influenza-negative hospitalized infants with ARI/FI enrolled during the months that cases were also enrolled; (2) in the women, those aged ≥18 years, hospitalized for cardiorespiratory illness during pregnancy or 42 days postpartum, who would have been eligible to receive IIV during pregnancy were enrolled; cases were those with PCR-confirmed influenza, and controls those who tested negative.
VE was calculated as 1 – odds ratio [ratio of odds of vaccination among cases to the odds of vaccination among controls] × 100% using logistic regression. Both VE estimates were adjusted for HIV infection (maternal analysis) or HIV exposure (infant analysis) and any other variables that were significantly associated with the exposure (vaccination) or the outcome (influenza infection) on univariate analysis. Other potential confounder variables were evaluated by assessing if they altered the VE estimates; however, none changed the adjusted VE by >10% and thus were not included in the final model. A variable was created to define influenza season timing at hospital admission by dividing each influenza season into early, middle, and late season terciles of equal duration; for the infant analysis, an “outside of season” period was also considered. The South African influenza seasons were defined using data from the influenza sentinel surveillance from the National Institute for Communicable Diseases [24]. Descriptive statistics were used to characterize cases and controls and the proportion of women who received IIV.
In the primary infant analysis, VE was estimated for hospitalizations associated with any PCR-confirmed influenza illness from birth until 6 months of age. Subgroup analyses were conducted by HIV infection status of the mother (ie, HIV exposure), restricted to infants born at term (defined as infants born at ≥37 weeks’ gestation or with a birthweight ≥2500 g) or to infants admitted at <3 months of age. For the maternal analysis, VE was estimated for hospitalizations occurring during the influenza seasons associated with any PCR-confirmed influenza illness at any stage during pregnancy and up until 42 days postpartum; subgroup analyses were conducted by HIV infection status. Both infant and maternal analyses were also done to estimate VE limited to viruses matched with the vaccine strains for each year. Odds ratios were considered statistically significant when the 95% CIs did not overlap 1.0. Analyses were performed with Stata software, version 13.1 (StataCorp, College Station, Texas).
Patient Consent Statement
The study was approved by the Human Research Ethics Committee of the University of the Witwatersrand (M140826) and University of Cape Town (835/2014) and conducted in accordance with Good Clinical Practice guidelines. The Centers for Disease Control and Prevention Institutional Review Board (IRB) relied on the local IRB review (#6746, 45 Code of Federal Regulations [CFR] part 46; 21 CFR part 56). Written informed consent was obtained from the infants’ legal guardian or the hospitalized women.
RESULTS
Each year, the vaccination of pregnant women started as soon as IIV was available in the country. The timing of vaccine availability was 22 April–31 July 2015; 4 April–16 August 2016; 3 April–16 August 2017; and 13 March–29 June 2018.
Infant Analyses
Overall, 3484 infants aged <6 months born to women eligible to have received IIV during pregnancy and who were born at least 14 days after vaccines were available were hospitalized with ARI/FI at the participating hospitals and enrolled in the study. All the infants were tested by PCR for influenza virus, of whom 83 (2.4%) had a positive result. Maternal vaccination status was unavailable for 3 of the 83 (3.6%) infants and, for another 9 (10.8%), their mothers were vaccinated <14 days before delivery. Among all the influenza-negative infants enrolled, 421 were randomly selected, of whom 25 (5.9%) were excluded because maternal vaccination status was unavailable and 25 because their mothers were vaccinated <14 days before delivery. Accordingly, 71 influenza-positive and 371 influenza-negative infants were included in the analysis (Supplementary Figure 1).
Infant cases were older than controls (mean age, 66.2 vs 46.2 days; P < .001), whereas all other demographic characteristics were similar between the 2 groups (Table 1).
Table 1.
Characteristics of Hospitalized Infants Born to Mothers With Known Influenza Vaccination Status Who Tested Influenza Polymerase Chain Reaction Positive and Randomly Selected Control Infants Testing Influenza Negative—South Africa, 2015–2018
| Characteristic | Influenza-Positive Cases (n = 71) | Influenza-Negative Controls (n = 371) | P Value |
|---|---|---|---|
| Age at admission | |||
| Mean age, d (SD) | 66.2 (38.4) | 46.2 (37.2) | <.001 |
| <3 mo | 54 (76.1) | 316 (85.2) | .057 |
| 3–5 mo | 17 (23.9) | 55 (14.8) | |
| HIV exposure | |||
| Exposed | 16/70 (22.9) | 67 (18.1) | .35 |
| Unexposed | 54/70 (77.1) | 304 (81.9) | |
| Sex | |||
| Male | 41 (57.8) | 206 (55.5) | .73 |
| Female | 30 (42.3) | 165 (44.5) | |
| Race | |||
| Black African | 50 (70.4) | 269 (72.5) | .94 |
| South African Coloured | 19 (26.8) | 89 (24.0) | |
| White | 1 (1.4) | 8 (2.2) | |
| Asian | 1 (1.4) | 5 (1.4) | |
| Term (≥37 wk gestation) or normal birthweight (≥2500 g) | |||
| Yes | 57 (80.3) | 318 (85.7) | .24 |
| No | 14 (19.7) | 53 (14.3) | |
| Ever breastfed | |||
| Yes | 54/69 (78.3) | 301/364 (82.7) | .38 |
| No | 15/69 (21.7) | 63/364 (17.3) | |
| Study site | |||
| Chris Hani Baragwanath Academic Hospital, Johannesburg | 26 (36.6) | 159 (42.9) | .74 |
| Rahima Moosa Mother and Child Hospital, Johannesburg | 17 (23.9) | 77 (20.8) | |
| Red Cross Hospital, Cape Town | 26 (36.6) | 121 (32.6) | |
| Mitchell's Plain Hospital, Cape Town | 2 (2.8) | 14 (3.8) | |
| Antenatal clinic attended supplemented with vaccine | |||
| Yes | 42/64 (65.6) | 214/348 (61.5) | .53 |
| No | 22/64 (34.4) | 134/348 (38.5) | |
| Year of enrollment | |||
| 2015 | 12 (16.9) | 49 (13.2) | .58 |
| 2016 | 24 (33.8) | 111 (29.9) | |
| 2017 | 14 (19.7) | 98 (26.4) | |
| 2018 | 21 (29.6) | 113 (30.5) | |
| Period of influenza season | |||
| Early | 14 (19.7) | 66 (17.8) | .46 |
| Middle | 20 (28.2) | 118 (31.8) | |
| Late | 28 (39.4) | 129 (34.8) | |
| Outside | 9 (12.7) | 58 (15.6) | |
| Influenza vaccination during pregnancy (>13 d before delivery) | |||
| Yes | 19 (26.8) | 132 (35.6) | .15 |
| No | 52 (73.2) | 239 (64.4) |
Values are presented as No. (%) unless stated otherwise. The number of participants with available information is indicated (no./No.) if different from the total number of participants.
Abbreviations: HIV, human immunodeficiency virus; SD, standard deviation.
Overall, 151 (34.3%) infants were born to vaccinated mothers, and 291 (65.7%) were born to unvaccinated mothers. Infants born to vaccinated mothers compared with unvaccinated, were younger at hospital admission, and a higher percentage were born at term. Furthermore, a higher percentage of vaccinated mothers attended antenatal care at the clinics where vaccines were supplied by the study. Other significant differences between the vaccinated and unvaccinated group were year of enrollment and timing during the influenza season that hospitalization occurred (Supplementary Table 2).
Across study seasons and sites, 26.8% (19/71) of mothers of the infant cases were vaccinated during pregnancy compared with 35.6% (132/371) of the control infants, yielding an adjusted VE (aVE) of 29.0% (95% CI, −33.6% to 62.3%). Similar nonsignificant aVE point estimates were obtained when restricting the analysis to infants <3 months old or to infants born at term. When only vaccine-matched strains across the different years were included in the analysis (ie, excluding 4 influenza B unsubtyped cases, 18 influenza B/Victoria, and 6 influenza B/Yamagata), aVE was 65.2% (95% CI, 11.7%–86.3%), with similar VE in infants <3 months old or born at term (Table 2).
Table 2.
Effectiveness of Influenza Vaccine Administered in Pregnancy Against Influenza-Confirmed Hospitalization in Infants <6 Months of Age—South Africa, 2015–2018
| Characteristic | Influenza Positive, No. Vaccinated/Total No. |
Influenza Negative, No. Vaccinated/Total No. |
Unadjusted VE, % (95% CI) |
Adjusted VE, % (95% CI)a |
|---|---|---|---|---|
| All infants | ||||
| Overall | 19/71 | 132/371 | 33.8 (−16.6 to 62.5) | 29.0 (−33.6 to 62.3) |
| Born at term | 16/57 | 122/318 | 37.3 (−16.6 to 66.3) | 38.4 (−23.5 to 69.3) |
| Admitted to hospital at <3 mo of age | 14/54 | 120/316 | 42.8 (−9.5 to 70.1) | 42.8 (−17.0 to 72.0) |
| Vaccine-matched strains | 7/43 | 132/371 | 64.8 (18.7–84.8) | 65.2 (11.7–86.3) |
| Born at term | 6/35 | 122/318 | 66.8 (17.6–86.6) | 69.3 (15.1–88.9) |
| Admitted to hospital at <3 mo of age | 6/36 | 120/316 | 67.3 (19.2–86.8) | 70.6 (21.0–89.1) |
| HIV unexposed | ||||
| Overall | 15/54 | 111/304 | 33.1 (−26.8 to 64.7) | 24.1 (−55.1 to 62.8) |
| Born at term | 12/41 | 104/267 | 35.1 (−32.7 to 68.3) | 34.6 (−45.9 to 70.6) |
| Admitted to hospital at <3 mo of age | 10/39 | 100/259 | 45.2 (−17.4 to 74.4) | 43.8 (−29.3 to 75.6) |
| Vaccine-matched strains | 6/31 | 111/304 | 58.3 (−4.8 to 83.4) | 50.6 (−39.6 to 82.5) |
| Born at term | 5/23 | 104/267 | 56.5 (−20.8 to 84.3) | 55.8 (−35.2 to 85.6) |
| Admitted to hospital at <3 mo of age | 5/26 | 100/259 | 62.1 (−3.6 to 86.2) | 62.2 (−10.7 to 87.1) |
| HIV exposed | ||||
| Overall | 4/16 | 21/67 | 27.0 (−153.3 to 78.9) | 28.7 (−226.4 to 84.4) |
| Born at term | 4/15 | 18/51 | 33.3 (−139.9 to 81.5) | 58.5 (−125.6 to 92.4) |
| Admitted to hospital at <3 mo of age | 4/14 | 20/57 | 26.0 (−166.4 to 79.4) | 26.7 (−232.5 to 83.8) |
| Vaccine-matched strains | 1/11 | 21/67 | 78.1 (−82.4 to 97.4) | 87.9 (−48.5 to 99.0) |
| Born at term | 1/11 | 18/51 | 81.7 (−54.9 to 97.8) | 96.1 (10.0–99.8) |
| Admitted to hospital at <3 mo of age | 1/9 | 20/57 | 76.9 (−98.3 to 97.3) | 86.9 (−61.7 to 98.9) |
Values in bold denote significant vaccine effectiveness estimates. Abbreviations: CI, confidence interval; HIV, human immunodeficiency virus; VE, vaccine effectiveness.
Vaccine effectiveness overall adjusted for year of study, hospital, period of influenza season when hospitalization occurred, age at hospitalization, antenatal clinic attended by the mother being supplemented with vaccine (yes or no), born at term (yes or no), and HIV exposure status.
Among HIV-unexposed infants, the maternal vaccination coverage was 27.8% in the cases and 36.8% in controls, corresponding to an aVE of 24.1% (95% CI, −55.1% to 62.8%). Only 16 PCR-confirmed influenza virus infections were detected among HIV-exposed infants and 25% of their mothers were vaccinated compared to 31.4% of mothers from control infants, for an aVE of 28.7% (95% CI, −226.4% to 84.4%). Similar VE point estimates were obtained in both HIV exposure–stratified groups restricting the analyses to infants <3 months old or born at term. For only vaccine-matched strains, the aVE was 50.6% (95% CI, −39.6% to 82.5%) among HIV-unexposed infants and 87.9% (95% CI, −48.5% to 99.0%) for HIV-exposed infants (Table 2).
Maternal Analyses
During the study period, 415 pregnant or postpartum women eligible to have received IIV during pregnancy were hospitalized during the influenza seasons at least 14 days after IIV was available. Fifty-nine (14.2%) PCR-confirmed influenza cases were detected, among whom vaccination status was unavailable for 3 (5.1%). Among the 356 influenza-negative controls, 3 (0.8%) had unknown vaccination status, 1 (0.3%) was vaccinated but date was missing, and 7 (2.0%) received vaccine <14 days before hospitalization. Therefore, 56 cases and 345 controls were included in the maternal analysis (Supplementary Figure 2). Cases and controls had similar characteristics (Table 3). Overall, 148 (36.9%) women received IIV before admission; vaccinated and unvaccinated women differed by admitting hospital, pregnancy trimester at admission, and year of hospitalization (Supplementary Table 3).
Table 3.
Characteristics of Hospitalized Pregnant and Postpartum Women With Known Influenza Vaccination Status Who Tested Influenza Positive or Negative by Polymerase Chain Reaction—South Africa, 2015–2018
| Characteristic | Influenza-Positive Cases (n = 56) | Influenza-Negative Controls (n = 345) | P Value |
|---|---|---|---|
| Age at admission, y | |||
| Mean (SD) | 30.1 (6.6) | 29.6 (6.3) | .60 |
| 18–24 | 14 (25.0) | 86 (25.0) | .99 |
| 25–34 | 27 (48.2) | 167 (48.6) | |
| ≥35 | 15 (26.8) | 91 (26.5) | |
| HIV infection | |||
| Infected | 28 (50.0) | 148/344 (43.0) | .33 |
| Uninfected | 28 (50.0) | 196/344 (57.0) | |
| Race | |||
| Black African | 40 (71.4) | 267 (77.4) | .67 |
| South African Coloured | 15 (26.8) | 69 (20.0) | |
| White | 1 (1.8) | 6 (1.7) | |
| Asian | 0 (0) | 3 (0.9) | |
| Pregnancy trimester at admission | |||
| First | 1/55 (1.8) | 8/340 (2.4) | .54 |
| Second | 13/55 (23.6) | 77/340 (22.7) | |
| Third | 40/55 (72.7) | 231/340 (67.9) | |
| Postpartum | 1/55 (1.7) | 24/340 (7.1) | |
| Study site | |||
| Chris Hani Baragwanath Academic Hospital, Johannesburg | 13 (23.2) | 96 (27.8) | .81 |
| Rahima Moosa Mother and Child Hospital, Johannesburg | 21 (37.5) | 123 (35.7) | |
| Helen Joseph Hospital, Johannesburg | 2 (3.6) | 12 (3.5) | |
| Groote-Schuur Hospital, Cape Town | 5 (8.9) | 34 (9.9) | |
| Mowbray Hospital, Cape Town | 10 (18.9) | 40 (11.6) | |
| Mitchell's Plain Hospital, Cape Town | 5 (8.9) | 40 (11.6) | |
| Comorbidity | |||
| Yes | 13 (23.2) | 59 (17.1) | .27 |
| No | 43 (76.8) | 286 (82.9) | |
| Smoker | |||
| Yes | 4 (7.1) | 31 (9.0) | .80 |
| No | 52 (92.9) | 314 (91.0) | |
| Alcohol consumption | |||
| Yes | 2 (3.6) | 16 (4.6) | .99 |
| No | 54 (96.4) | 329 (95.4) | |
| Period of influenza season | |||
| Early | 24 (42.9) | 134 (38.8) | .72 |
| Middle | 20 (35.7) | 120 (34.8) | |
| Late | 12 (21.4) | 91 (26.4) | |
| Year of enrollment | |||
| 2015 | 4 (7.1) | 34 (9.9) | .56 |
| 2016 | 19 (33.9) | 91 (26.4) | |
| 2017 | 15 (26.8) | 85 (24.6) | |
| 2018 | 18 (32.1) | 135 (39.1) | |
| Influenza vaccination during pregnancy >13 d before admission | |||
| Yes | 16 (28.6) | 132 (38.3) | .16 |
| No | 40 (71.4) | 213 (61.7) |
Values are presented as No. (%) unless stated otherwise. The number of participants with available information is indicated (no./No.) if different from the total number of participants.
Abbreviations: HIV, human immunodeficiency virus; SD, standard deviation.
Overall, 28.6% of the cases were vaccinated compared with 38.3% of the controls, for an aVE of 46.9% (95% CI, −2.8% to 72.5%). When analysis was stratified by HIV infection status, HIV-uninfected women had a vaccination coverage of 10.7% among cases and 40.3% among controls, resulting in an aVE of 82.8% (95% CI, 40.7%–95.0%). In WLWH, vaccination coverage was 46.4% and 35.8% among cases and controls, respectively (aVE, −32.5% [95% CI, −208.7% to 43.1%]). Similar results were obtained when restricted to vaccine-matched strains (Table 4).
Table 4.
Effectiveness of Influenza Vaccine Administered in Pregnancy Against Maternal Influenza-Confirmed Hospitalization—South Africa, 2015–2018
| Participants | Influenza Positive, No. Vaccinated/Total No. |
Influenza Negative, No. Vaccinated/Total No. |
Unadjusted VE, % (95% CI) |
Adjusted VE, % (95% CI)a |
|---|---|---|---|---|
| All mothers | ||||
| Overall | 16/56 | 132/345 | 35.5 (−19.9 to 65.2) | 46.9 (−2.8 to 72.5) |
| Vaccine-matched strains | 13/43 | 132/345 | 30.1 (−38.9 to 64.8) | 44.0 (−17.4 to 73.3) |
| Without HIV | ||||
| Overall | 3/28 | 79/196 | 82.2 (39.1–94.8) | 82.8 (40.7–95.0) |
| Vaccine-matched strains | 2/22 | 79/196 | 85.2 (34.9–96.6) | 85.8 (36.7–96.8) |
| Living with HIV | ||||
| Overall | 13/28 | 53/148 | −55.3 (−251.0 to 31.2) | −32.5 (−208.7 to 43.1) |
| Vaccine-matched strains | 11/21 | 53/148 | −97.2 (−394.7 to 21.4) | −78.9 (−370.5 to 32.0) |
Values in bold denote significant vaccine effectiveness estimates. Abbreviations: CI, confidence interval; HIV, human immunodeficiency virus; VE, vaccine effectiveness.
VE adjusted for year of study, hospital, pregnancy trimester or postpartum period when hospitalization occurred, and HIV infection status for all mothers.
Circulating Strains
The percentage of influenza strains detected among study participants matched to the annual vaccines’ formulations varied from 38.7% in 2018, when 2 distinct peaks of A(H1N1)pdm09 and influenza B/Victoria (vaccine mismatch) were detected in the country, to 93% in 2016 when there was a good match between influenza B and A(H3N2) co-circulating strains and the vaccine strains (Table 5).
Table 5.
Influenza Strains Detected During the Study Period in Infants and Women
| Influenza Strain | 2015 | 2016 | 2017 | 2018 | Overall |
|---|---|---|---|---|---|
| A(H1N1)pdm09 | 2 | 12 | 1 | 15 | 29 |
| A(H3N2) | 1 | 12 | 19 | … | 32 |
| B/Victoria | 2 | 16 | … | 24 | 42 |
| B/Yamagata | 9 | … | 9 | … | 18 |
| B (Unsubtyped) | 2 | 3 | … | … | 5 |
| Vaccine-matched strains | 11 (68.8%) | 40 (93.0%) | 20 (69.0%) | 15 (38.5%) | 86 (67.7%) |
Bolded values indicate viruses similar to the vaccine viruses.
DISCUSSION
In our study covering 4 consecutive influenza seasons, influenza vaccination during pregnancy had an estimated effectiveness of 65% against influenza-associated hospital admissions in young infants, although VE was only demonstrated against influenza viruses considered to be vaccine matches. Overall, 84% of all enrolled infants were aged <3 months at hospital admission, and therefore the VE point estimates were similar considering all ages or restricting to infants aged <3 months. In a stratified analysis by HIV exposure status, although the VE point estimates for vaccine-matched strains were 58% for HIV-unexposed infants and 78% for HIV-exposed infants, these were nonsignificant, due to the small number of cases that only provided power of 48% and 27%, respectively. During the study period, surveillance was also performed for influenza-associated hospitalizations among women. Among HIV-uninfected women, high VE estimates were detected either for all influenza (82%) and for vaccine-matched strains only (85%). In this study, maternal influenza vaccination did not demonstrate effectiveness against hospitalization in WLWH.
In previous RCTs over 2 influenza seasons in South Africa, the efficacy of influenza vaccination during pregnancy against any PCR-confirmed influenza illness was similar among HIV-uninfected women (50.4% [95% CI, 14.5%–71.2%]) and WLWH (57.7% [95% CI, .2%–82.1%]), notwithstanding WLWH having lower humoral immune responses following vaccination [7–9]. The efficacy of influenza vaccination against confirmed influenza illness was also 75.5% (95% CI, 9.2%–95.6%) among nonpregnant South African adults with HIV [25]. We are not aware of any study reporting on the efficacy of influenza vaccination specifically against influenza-associated hospitalization among people with HIV. Residual confounding inherent to observational studies may, however, explain the differences in the current results with the RCTs.
Throughout the entire study, both influenza A and B viruses circulated in South Africa. Genetic data showed that during the study period, most of the circulating A(H1N1)pdm09 viruses were similar to the ones included in the seasonal vaccines [20–23]. For A(H3N2), using national sequencing information, for each study year but 2015, there was good concordance between the circulating and vaccine viruses. In 2015 the lineage 3C.3a was included in the vaccine, but circulating strains were in the 3C.2a lineage and therefore the A(H3N2) cases in 2015 were considered not similar to vaccine viruses. For the influenza B lineages, the viruses identified in study participants during 2017 and 2018 were not of the same lineage that was contained in the trivalent vaccine. It has been suggested that trivalent IIV offers cross-protection against nonvaccine B lineages [26, 27]; in our study, however, where most of the influenza B vaccine-unmatched viruses were detected in the infants, a protective effect in the infants was only detected for vaccine matches. This result is different from the vaccine efficacy among HIV-unexposed infants in the South African RCT, which was similar either including all PCR-confirmed influenza episodes (48.8% [95% CI, 11.6%–70.40%]) or excluding the 17 cases of nonvaccine B lineage (48.2% [95% CI, −.8% to 73.3%]) [1]. Our VE of 65% among the infants is in line with previous reports from Europe and the US where estimates of 45%–92% were reported [3, 10–15]. A study from England that measured the effectiveness of maternal influenza vaccination in preventing influenza-associated hospitalizations in infants aged <6 months over 2 consecutive seasons of 2013–2014 (dominated by A[H1N1]pdm09) and 2014–2015 (dominated by a drifted A[H3N2] strain) reported an overall VE of 64% (95% CI, 5%–87%) in 2013–2014 and 50% (95% CI, 8%–73%) in 2014–2015, with a similar estimate for 2014–2015 (58% [95% CI, 7%–81%]) if restricted to infants infected with the dominant drifted A(H3N2) strain [15].
Our VE estimates among HIV-uninfected women are higher than previously reported in a study across 4 countries from 2010 to 2016 (40% [95% CI, 12%–59%]), although in that study only 16% of all hospitalized women were vaccinated [16].
From 2015 to 2018, our study supplemented the national influenza vaccination program at selected antenatal clinics aiming at increasing vaccination rates to at least 50%. Although the vaccination campaigns were very successful and >75% of the women who received care at the selected clinics were vaccinated during the campaigns [17], the vaccination coverage among the hospitalized study participants, many of whom sought antenatal care at clinics which did not have supplemental vaccine supply, was <36%. This low vaccination coverage impacted the power of our analyses and might be the result of several factors. First, although analyses were restricted to those who were eligible to have received IIV during pregnancy, it is possible that not all participants had antenatal visits during the period that vaccines were available, even though our campaigns lasted for 3–4 months, and South African national estimates for 2016 revealed that 76% of pregnant women attended ≥4 antenatal care visits and that 47% of women had their first visit during the first pregnancy trimester [28]. Second, not all women attended antenatal care at clinics supplemented with study vaccine and, as shown in Supplementary Table 2, only 18% of the women attending other clinics were vaccinated. Third, although the vaccination campaigns aimed to deliver vaccines before the start of the influenza season, the actual timing of vaccination was determined by the availability of vaccines at the selected clinics, and as previously reported across the study years, approximately only 52% of the vaccines were administered prior to the start of the influenza seasons [17]. Vaccination campaigns aimed at reducing the burden of influenza-associated illness, especially among infants, need to consider how quickly vaccination can be implemented once vaccines are available to maximize the overlap between vaccination opportunity and the risk of influenza infection in the population.
Limitations of our study include that the vaccination coverage among study participants was lower than initially anticipated, leading to a decrease in the power of the analyses. We tried to systematically document the vaccination status of the women attending the antenatal clinics selected for vaccine supplementation, and in the analyses, women were considered vaccinated only if written evidence was available; however, women could have been vaccinated and if this information was not recorded in their antenatal cards or in the study registers, they would be misclassified as unvaccinated. This misclassification would likely have been nondifferential, which would potentially result in an underestimation of the VE. Since written informed consent was required for participation, not every eligible patient was included in the study; however, differences in participation should have been similar according to influenza infection and vaccination status. Overall, as this is an observational study, there could be residual confounding that we were unable to account for. A further limitation is that we were unable to investigate VE in WLWH stratified by their degree of immunosuppression, as these data were not available.
Our study provides additional evidence for the benefit of maternal influenza vaccination to prevent severe disease in infants and among HIV-uninfected women. The effectiveness in WLWH may depend on the degree of immunosuppression, but we were unable to assess that, and further studies are needed.
Supplementary Material
Contributor Information
Marta C Nunes, South African Medical Research Council, Faculty of Health Sciences, Vaccines and Infectious Diseases Analytics Research Unit, University of the Witwatersrand, Johannesburg, South Africa; Department of Science and Technology/National Research Foundation, Faculty of Health Sciences, South African Research Chair Initiative in Vaccine Preventable Diseases, University of the Witwatersrand, Johannesburg, South Africa.
Sibongile Walaza, Centre for Respiratory Diseases and Meningitis, National Institute for Communicable Diseases, Johannesburg, South Africa; School of Public Health, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa.
Susan Meiring, Centre for Respiratory Diseases and Meningitis, National Institute for Communicable Diseases, Johannesburg, South Africa; School of Public Health, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa; Division of Public Health Surveillance and Response, National Institute for Communicable Diseases, Johannesburg, South Africa.
Heather J Zar, Department of Paediatrics and Child Health, and South African Medical Research Council Unit on Child and Adolescent Health, University of Cape Town, Cape Town, South Africa.
Gary Reubenson, Department of Paediatrics and Child Health, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa.
Meredith McMorrow, Influenza Division, Centers for Disease Control and Prevention, Atlanta, Georgia, USA.
Stefano Tempia, School of Public Health, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa; Influenza Division, Centers for Disease Control and Prevention, Atlanta, Georgia, USA.
Liza Rossi, Centre for Respiratory Diseases and Meningitis, National Institute for Communicable Diseases, Johannesburg, South Africa.
Raphaela Itzikowitz, Centre for Respiratory Diseases and Meningitis, National Institute for Communicable Diseases, Johannesburg, South Africa.
Kate Bishop, Centre for Respiratory Diseases and Meningitis, National Institute for Communicable Diseases, Johannesburg, South Africa.
Azwifarwi Mathunjwa, Centre for Respiratory Diseases and Meningitis, National Institute for Communicable Diseases, Johannesburg, South Africa.
Amy Wise, Department of Obstetrics and Gynaecology, Rahima Moosa Mother and Child Hospital, University of the Witwatersrand, Johannesburg, South Africa.
Florette K Treurnicht, Centre for Respiratory Diseases and Meningitis, National Institute for Communicable Diseases, Johannesburg, South Africa.
Orienka Hellferscee, Centre for Respiratory Diseases and Meningitis, National Institute for Communicable Diseases, Johannesburg, South Africa; School of Public Health, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa.
Matt Laubscher, South African Medical Research Council, Faculty of Health Sciences, Vaccines and Infectious Diseases Analytics Research Unit, University of the Witwatersrand, Johannesburg, South Africa; Department of Science and Technology/National Research Foundation, Faculty of Health Sciences, South African Research Chair Initiative in Vaccine Preventable Diseases, University of the Witwatersrand, Johannesburg, South Africa.
Natali Serafin, South African Medical Research Council, Faculty of Health Sciences, Vaccines and Infectious Diseases Analytics Research Unit, University of the Witwatersrand, Johannesburg, South Africa; Department of Science and Technology/National Research Foundation, Faculty of Health Sciences, South African Research Chair Initiative in Vaccine Preventable Diseases, University of the Witwatersrand, Johannesburg, South Africa.
Clare L Cutland, South African Medical Research Council, Faculty of Health Sciences, Vaccines and Infectious Diseases Analytics Research Unit, University of the Witwatersrand, Johannesburg, South Africa; Department of Science and Technology/National Research Foundation, Faculty of Health Sciences, South African Research Chair Initiative in Vaccine Preventable Diseases, University of the Witwatersrand, Johannesburg, South Africa; African Leadership in Vaccinology Expertise, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa.
Shabir A Madhi, South African Medical Research Council, Faculty of Health Sciences, Vaccines and Infectious Diseases Analytics Research Unit, University of the Witwatersrand, Johannesburg, South Africa; Department of Science and Technology/National Research Foundation, Faculty of Health Sciences, South African Research Chair Initiative in Vaccine Preventable Diseases, University of the Witwatersrand, Johannesburg, South Africa; African Leadership in Vaccinology Expertise, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa.
Cheryl Cohen, Centre for Respiratory Diseases and Meningitis, National Institute for Communicable Diseases, Johannesburg, South Africa; School of Public Health, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa.
Supplementary Data
Supplementary materials are available at Open Forum Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.
Notes
Acknowledgments. The authors thank the study participants, as well as the clinic, laboratory, and data teams of the Vaccines and Infectious Diseases Analytics Research Unit of the University of the Witwatersrand (Wits-VIDA) and the National Institute for Communicable Diseases. The authors also acknowledge the expert advice provided by the study scientific advisory committee members: Haroon Saloojee, Danuta Skowronski, Gaston de Serres, Jennifer Verani, and Joseph Bresee.
Disclaimer. The funders had no role in the design, analysis, or interpretation of data.
Financial support. This work was supported by the National Institute for Communicable Diseases of the National Health Laboratory Service; the US Centers for Disease Control and Prevention (CDC) (cooperative agreement number 5U51IP000155); and the Bill and Melinda Gates Foundation (BMGF) (grant number OPP1118349). There was also partial support from the Department of Science and Technology and National Research Foundation: South African Research Chair Initiative in Vaccine Preventable Diseases, and the South African Medical Research Council: Wits-VIDA Research Unit.
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.
References
- 1. Madhi SA, Cutland CL, Kuwanda L, et al. . Influenza vaccination of pregnant women and protection of their infants. N Engl J Med 2014; 371:918–31. [DOI] [PubMed] [Google Scholar]
- 2. Tapia MD, Sow SO, Tamboura B, et al. . Maternal immunisation with trivalent inactivated influenza vaccine for prevention of influenza in infants in Mali: a prospective, active-controlled, observer-blind, randomised phase 4 trial. Lancet Infect Dis 2016; 16:1026–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Nunes MC, Madhi SA. Influenza vaccination during pregnancy for prevention of influenza confirmed illness in the infants: a systematic review and meta-analysis. Hum Vaccin Immunother 2018; 14:758–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Steinhoff MC, Katz J, Englund JA, et al. . Year-round influenza immunisation during pregnancy in Nepal: a phase 4, randomised, placebo-controlled trial. Lancet Infect Dis 2017; 17:981–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Steinhoff MC, Omer SB, Roy E, et al. . Influenza immunization in pregnancy—antibody responses in mothers and infants. N Engl J Med 2010; 362:1644–6. [DOI] [PubMed] [Google Scholar]
- 6. Nunes MC, Cutland CL, Jones S, et al. . Duration of infant protection against influenza illness conferred by maternal immunization: secondary analysis of a randomized clinical trial. JAMA Pediatr 2016; 170:840–7. [DOI] [PubMed] [Google Scholar]
- 7. Nunes MC, Cutland CL, Dighero B, et al. . Kinetics of hemagglutination-inhibiting antibodies following maternal influenza vaccination among mothers with and those without HIV infection and their infants. J Infect Dis 2015; 212:1976–87. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Nunes MC, Weinberg A, Cutland CL, et al. . Neutralization and hemagglutination-inhibition antibodies following influenza vaccination of HIV-infected and HIV-uninfected pregnant women. PLoS One 2018; 13:e0210124. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Dhar N, Kwatra G, Nunes MC, et al. . Hemagglutinin-stalk antibody responses following trivalent inactivated influenza vaccine immunization of pregnant women and association with protection from influenza virus illness. Clin Infect Dis 2020; 71:1072–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Benowitz I, Esposito DB, Gracey KD, Shapiro ED, Vázquez M. Influenza vaccine given to pregnant women reduces hospitalization due to influenza in their infants. Clin Infect Dis 2010; 51:1355–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Dabrera G, Zhao H, Andrews N, et al. . Effectiveness of seasonal influenza vaccination during pregnancy in preventing influenza infection in infants, England, 2013/14. Euro Surveill 2014; 19:20959. [DOI] [PubMed] [Google Scholar]
- 12. Poehling KA, Szilagyi PG, Staat MA, et al. . Impact of maternal immunization on influenza hospitalizations in infants. Am J Obstet Gynecol 2011; 204(6 Suppl 1):S141–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Shakib JH, Korgenski K, Presson AP, et al. . Influenza in infants born to women vaccinated during pregnancy. Pediatrics 2016; 137:e20152360. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Molgaard-Nielsen D, Fischer TK, Krause TG, Hviid A. Effectiveness of maternal immunization with trivalent inactivated influenza vaccine in pregnant women and their infants. J Intern Med 2019; 286:469–80. [DOI] [PubMed] [Google Scholar]
- 15. Walker JL, Zhao H, Dabrera G, et al. . Assessment of effectiveness of seasonal influenza vaccination during pregnancy in preventing influenza infection in infants in England, 2013–2014 and 2014–2015. J Infect Dis 2020; 221:16–20. [DOI] [PubMed] [Google Scholar]
- 16. Thompson MG, Kwong JC, Regan AK, et al. . Influenza vaccine effectiveness in preventing influenza-associated hospitalizations during pregnancy: a multi-country retrospective test negative design study, 2010–2016. Clin Infect Dis 2019; 68:1444–53. [DOI] [PubMed] [Google Scholar]
- 17. Bishop K, McMorrow M, Meiring S, et al. . An evaluation of an influenza vaccination campaign targeting pregnant women in 27 clinics in two provinces of South Africa, 2015–2018. BMC Health Serv Res 2021; 21:941. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Jernigan DB, Lindstrom SL, Johnson JR, et al. . Detecting 2009 pandemic influenza A (H1N1) virus infection: availability of diagnostic testing led to rapid pandemic response. Clin Infect Dis 2011; 52(Suppl 1):S36–43. [DOI] [PubMed] [Google Scholar]
- 19. McAnerney JM, Walaza S, Tempia S, et al. . Estimating vaccine effectiveness in preventing laboratory-confirmed influenza in outpatient settings in South Africa, 2015. Influenza Other Respir Viruses 2017; 11:177–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.The Crick Worldwide Influenza Centre, UK. https://www.crick.ac.uk/sites/default/files/2018-07/september_2016_interim_report.pdf. Accessed 27 March 2022.
- 21.The Crick Worldwide Influenza Centre, UK. https://www.crick.ac.uk/sites/default/files/2018-10/September%202018%20interim%20report_opt.pdf. Accessed 27 March 2022.
- 22.The Crick Worldwide Influenza Centre, UK. https://www.crick.ac.uk/sites/default/files/2018-07/crick_sh2017_vcm_report_to_post.pdf. Accessed 27 March 2022.
- 23.The Crick Worldwide Influenza Centre, UK. https://www.crick.ac.uk/sites/default/files/2018-07/crick_sep2015_vcm_report_to_post.pdf. Accessed 27 March 2022.
- 24. National Institute for Communicable Diseases, South Africa . About us. https://www.nicd.ac.za/about-us/. Accessed 27 March 2022.
- 25. Madhi SA, Maskew M, Koen A, et al. . Trivalent inactivated influenza vaccine in African adults infected with human immunodeficient virus: double blind, randomized clinical trial of efficacy, immunogenicity, and safety. Clin Infect Dis 2011; 52:128–37. [DOI] [PubMed] [Google Scholar]
- 26. McLean HQ, Thompson MG, Sundaram ME, et al. . Influenza vaccine effectiveness in the United States during 2012–2013: variable protection by age and virus type. J Infect Dis 2015; 211:1529–40. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Skowronski DM, Janjua NZ, Sabaiduc S, et al. . Influenza A/subtype and B/lineage effectiveness estimates for the 2011–2012 trivalent vaccine: cross-season and cross-lineage protection with unchanged vaccine. J Infect Dis 2014; 210:126–37. [DOI] [PubMed] [Google Scholar]
- 28. http://www.statssa.gov.za/?p=13100.
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.