Since January 2020 Elsevier has created a COVID-19 resource centre with

free information in English and Mandarin on the novel coronavirus COVID-

19. The COVID-19 resource centre is hosted on Elsevier Connect, the
company's public news and information website.

Elsevier hereby grants permission to make all its COVID-19-related

research that is available on the COVID-19 resource centre - including this
research content - immediately available in PubMed Central and other
publicly funded repositories, such as the WHO COVID database with rights
for unrestricted research re-use and analyses in any form or by any means
with acknowledgement of the original source. These permissions are
granted for free by Elsevier for as long as the COVID-19 resource centre
remains active.
 Vaccine 38 (2020) 350–354

Contents lists available at ScienceDirect

Vaccine
journal homepage: [Link]/locate/vaccine

Influenza vaccination and respiratory virus interference among

Department of Defense personnel during the 2017–2018 influenza
season
Greg G. Wolff
Armed Forces Health Surveillance Branch Air Force Satellite, 2510 5th Street, Bldg 840, Wright-Patterson AFB, OH 45433, United States

a r t i c l e i n f o a b s t r a c t

Article history: Purpose: Receiving influenza vaccination may increase the risk of other respiratory viruses, a phe-
Received 20 June 2019 nomenon known as virus interference. Test-negative study designs are often utilized to calculate influ-
Received in revised form 30 September enza vaccine effectiveness. The virus interference phenomenon goes against the basic assumption of
2019
the test-negative vaccine effectiveness study that vaccination does not change the risk of infection with
Accepted 1 October 2019
Available online 10 October 2019
other respiratory illness, thus potentially biasing vaccine effectiveness results in the positive direction.
This study aimed to investigate virus interference by comparing respiratory virus status among
Department of Defense personnel based on their influenza vaccination status. Furthermore, individual
Keywords:
Influenza vaccine
respiratory viruses and their association with influenza vaccination were examined.
Virus interference Results: We compared vaccination status of 2880 people with non-influenza respiratory viruses to 3240
Department of Defense people with pan-negative results. Comparing vaccinated to non-vaccinated patients, the adjusted odds
Respiratory illness ratio for non-flu viruses was 0.97 (95% confidence interval (CI): 0.86, 1.09; p = 0.60). Additionally, the vac-
cination status of 3349 cases of influenza were compared to three different control groups: all controls
(N = 6120), non-influenza positive controls (N = 2880), and pan-negative controls (N = 3240). The
adjusted ORs for the comparisons among the three control groups did not vary much (range: 0.46–0.51).
Conclusions: Receipt of influenza vaccination was not associated with virus interference among our pop-
ulation. Examining virus interference by specific respiratory viruses showed mixed results. Vaccine
derived virus interference was significantly associated with coronavirus and human metapneumovirus;
however, significant protection with vaccination was associated not only with most influenza viruses,
but also parainfluenza, RSV, and non-influenza virus coinfections.
Published by Elsevier Ltd.

1. Introduction vaccine was shown to be effective at reducing the burden of sea-

sonal influenza in the United States [2–6].
The influenza pandemic of 1918–1919, which contributed to an While influenza vaccination offers protection against influenza,
estimated 50 million deaths worldwide, stimulated interest in natural influenza infection may reduce the risk of non-influenza
influenza vaccine research [1]. Twenty years after the pandemic respiratory viruses by providing temporary, non-specific immunity
began, the first influenza vaccine was administered to US soldiers against these viruses [7,8]. On the other hand, recently published
in 1938 [1]. From the 2010–2011 influenza season to the 2017– studies have described the phenomenon of vaccine-associated
2018 season, excluding for the 2014–2015 season, the influenza virus interference; that is, vaccinated individuals may be at
increased risk for other respiratory viruses because they do not
Abbreviations: DoD, Department of Defense; DoDGRS, The Department of receive the non-specific immunity associated with natural infec-
Defense Global Respiratory Pathogen Surveillance Program; GEIS, Global Emerging tion [7–10]. There has been limited evidence that the influenza
Infections Surveillance and Response System; DHA/AFHSB, AF, The Defense Health vaccine may actually be associated with the virus interference pro-
Agency/Armed Forces Health Surveillance Branch, Air Force Satellite Cell; USAFSAM, cess [8,11]. Other studies have found no association between influ-
United States Air Force School of Aerospace Medicine; LRMC, Landstuhl Regional
enza vaccination and increased respiratory virus risk [10,12].
Medical Center; RSV, Respiratory Syncytial Virus; AFCITA, Air Force Complete
Immunization Tracking Application; OR, Odds Ratio; 95% CI, 95% Confidence The purpose of this study is to add to the general knowledge of
Interval. influenza vaccine-related virus interference by comparing rates of
E-mail address: [Link].3@[Link]

[Link]
0264-410X/Published by Elsevier Ltd.
 G.G. Wolff / Vaccine 38 (2020) 350–354 351

non-influenza respiratory viruses to negative laboratory tests, and NC). Basic descriptive epidemiology was utilized to obtain counts
comparing vaccination status of influenza positive cases to controls and rates of outcomes by sex, military beneficiary category, age
among Department of Defense (DoD) personnel. The DoD provides group, disease status, seasonality of illness, and vaccination status.
a unique population for vaccination studies as mandatory vaccina- In order to determine if virus interference was associated with
tion against influenza is required by the DoD for all Active Duty influenza vaccination in the military beneficiary population, odds
and Reserve Component personnel [13]. This study aims to exam- ratios and confidence intervals were calculated utilizing condi-
ine the relationship between specific respiratory viruses and influ- tional logistic regression to compare vaccination status from two
enza vaccination. The protocol for this study was reviewed and analyses. First, those with a viral respiratory illness other than
approved as exempt by the Air Force Research Laboratory Institu- influenza were compared to those with no pathogen detected
tional Review Board. (pan-negative). Next influenza positive cases were compared to
three different control groups. The first control group was com-
prised of all controls, specifically, individuals testing negative for
2. Materials and methods flu or positive for any respiratory virus other than flu. The second
control group consisted of only those who were positive for respi-
The Department of Defense Global Respiratory Pathogen ratory viruses other than influenza. Lastly, pan-negative controls
Surveillance Program (DoDGRS) is a DoD-wide program estab- were compared to influenza cases. Unadjusted and adjusted odds
lished by the Global Emerging Infections Surveillance and ratios were calculated for the overall population, the population
Response System (GEIS). The program was founded in 1997 as an with AFCITA records only, and the active duty only population dur-
influenza-only surveillance program. In the 2013–2014 influenza ing the influenza season for the comparison of other respiratory ill-
season the program added respiratory Film Array for flu negative nesses to pan-negatives, as well as all three case-control
samples and began identifying other respiratory pathogens. Start- comparisons. Adjusted odds ratios were calculated after modeling
ing in the 2017–2018 influenza season, the program added Lumi- variables in a nested logistic regression, keeping all variables with
nex Film Array capabilities to test for respiratory pathogens, and p < 0.20 and then adding them to a full logistic model. In the full
became known as DoDGRS. The Defense Health Agency/Armed logistic model, only variables that remained significant were
Forces Health Surveillance Branch – Air Force Satellite Cell (DHA/ included in the final adjusted model. Age group remained signifi-
AFHSB – AF) and United States Air Force School of Aerospace Med- cant in the overall population; age group and seasonality remained
icine (USAFSAM) manage the surveillance program that includes significant in the AFCITA confirmed vaccination group and the
global surveillance among DoD beneficiaries at 79 sentinel sites Active Duty population; and gender, age group, and seasonality
(including deployed locations) and many non-sentinel sites. Labo- all remained significant in all three of the case-control compar-
ratory testing completed at USAFSAM and Landstuhl Regional isons. Those respective variables that remained significant were
Medical Center (LRMC) included multiplex PCR respiratory patho- included in the adjustment for the odds ratio for the total season.
gen panels (including: adenovirus, Chlmydia pneumoniae, coron- Individual respiratory virus outcomes were also examined and
avirus, human bocavirus, human metapnumovirus, Mycoplasma stratified by vaccination status. Odds ratios, confidence intervals,
pneumoniae, parainfluenza, respiratory syncytial virus (RSV), rhi- and p-values were calculated to determine if individual respiratory
novirus/enterovirus, and co-infections) [14,15], viral culture viruses were associated with influenza vaccination.
detecting influenza and other respiratory viruses, and influenza
A/B subtyping via PCR [16,17]. Vaccination status was derived from
both the Air Force Complete Immunization Tracking Application 3. Results
(AFCITA), a United States Air Force database containing
vaccination-related data, and from surveys given to those submit- For the 2017–2018 influenza season, 4041 out of 11,943 speci-
ting respiratory samples. If the patient had an influenza vaccina- mens tested positive for influenza (33.8%) (Data not shown). There
tion record in AFCITA for the 2017–2018 influenza season, or were 3869 specimens identified as other respiratory pathogens
answered yes to being vaccinated during the season on their sur- (32.4%). The remaining 4033 specimens resulted as negative
vey, they were identified as vaccinated. Patients who were not vac- (33.8%) (Data not shown). Of the 11,943 specimens, 2474 (20.7%)
cinated for the season or who were vaccinated less than 14 days specimens were excluded from our population based on the exclu-
prior to specimen submittal were classified as unvaccinated. sionary criteria described in the Methods section, leaving a final
All people submitting a respiratory specimen to the DoDGRS for study population of 9469 unique people (Data not shown). The
the 2017–2018 influenza season were eligible for the study. The study population was predominantly male, Active Duty service
influenza season began 1 October 2017 and ended 29 September members, aged 18–35 years old (Tables 1 and 2). A majority of
2018. Those who submitted a sample and only tested positive for the study population was vaccinated (Tables 1 and 2). Most respi-
Chlamydia pneumoniae and/or Mycoplasma pneumoniae were ratory specimens were analyzed during the winter (December, Jan-
excluded because these illnesses are bacteriological in nature, not uary, and February) months (Tables 1 and 2).
viral. People with influenza and non-influenza coinfections were Those who tested positive for a respiratory virus other than
excluded because they could not be uniquely classified as either influenza had a similar breakdown for sex, vaccination status,
influenza or non-influenza respiratory virus. Individuals with mul- and season of illness when compared to pan-negatives (Table 1).
tiple specimens collected during the season were also removed The other respiratory positive group had more child beneficiaries,
from the study as they could have had multiple different viruses and was overall younger than the pan-negative group (Table 1).
over the season. Specimens where neither vaccination status could Examining demographic characteristics stratified by vaccination
be obtained via databases nor a questionnaire was completed were status, males were more likely to be vaccinated than females
excluded because vaccination status could not be confirmed. Sub- (Table 2). Active Duty members were more likely to be vaccinated
jects who were ill before receiving vaccination were excluded as than people with other beneficiary statuses (Table 2). The younger
vaccination status would therefore be unrelated to illness. Lastly, aged population was more likely to be unvaccinated when com-
those people for whom the laboratory rejected the specimen were pared to other age groups (Table 2). Composition of lab results
not included in the final study population. (Other respiratory virus, influenza, and no pathogen detected)
Data management and statistical analyses were conducted was distributed fairly evenly among the vaccinated population;
using SAS 9.4 and SAS Enterprise Guide 7.1 (SAS Institute Inc., Cary, however, unvaccinated people were more likely to have their
352 G.G. Wolff / Vaccine 38 (2020) 350–354

Table 1 significant (p = 0.60) (Table 3). Since self-reported vaccination sta-

Demographics by disease status. tus may not be accurate and may bias results, those with AFCITA
Respiratory Virus Positive No Pathogen Detected confirmed vaccination were examined exclusively. Those who
(n = 2880) (n = 3240) were vaccinated according AFCITA records had 5% lower unad-
N (column %) N (column %) justed odds (95% CI: 0.68, 1.34) of having other respiratory viruses
Sex compared to those who were unvaccinated (Table 3). Adjusting for
Male 1576 (54.7) 1770 (54.6) age group and seasonality increased the odds to 23% higher (95%
Female 1304 (45.3) 1470 (45.4)
CI: 0.86, 1.76) of having other respiratory viruses in the vaccinated
Beneficiary Category population (Table 3). Neither the unadjusted odds (data not
Active Duty 1119 (38.8) 1625 (50.1)
Child 1212 (42.1) 666 (20.6)
shown) nor the adjusted odds (p = 0.24) in the AFCITA population
Other 97 (3.4) 158 (4.9) were significant. Virus interference was also examined among
Retiree 103 (3.6) 244 (7.5) Active Duty only for the 2017–2018 season. Those who were vac-
Spouse 308 (10.7) 485 (15.0) cinated had slightly lower unadjusted odds (OR: 0.97, 95% CI: 0.73,
Unknown 41 (1.4) 62 (1.9)
1.29) of having other respiratory viruses compared to those who
Age Group were unvaccinated (Table 3). After adjusting for age and season,
0–17 1237 (42.9) 685 (21.1)
these odds increased to a 20% higher odds (95% CI: 0.89, 1.61) of
18–35 1074 (37.3) 1562 (48.2)
36+ 569 (19.8) 993 (30.7) having other respiratory viruses in the vaccinated population;
however, the unadjusted (data not shown) and adjusted
Vaccination Status
Vaccinated 2050 (71.2) 2441 (75.3) (p = 0.24) odds ratios were not statistically significant (Table 3).
Unvaccinated 830 (28.8) 799 (24.7) Both the unadjusted and adjusted odds of influenza were
Season significantly lower in the vaccinated population for all three of
Winter 1770 (61.4) 1921 (59.3) the control groups (Table 4). The adjusted ORs ranged from 0.46
Spring 716 (24.9) 827 (25.5) (pan-negative comparison) to 0.51 (non-influenza virus positive
Summer 69 (2.4) 122 (3.8) comparison) (Table 4). The 95% CI for the adjusted ORs of all three
Fall 325 (11.3) 370 (11.4)
comparison groups overlapped and no differences were detected
among each control group when compared to influenza cases.
The odds of testing positive for individual respiratory viruses by
Table 2
Demographics by vaccination status.
vaccination status were also examined (Table 5). The influenza
vaccine was sufficient at protecting all influenza virus results
Vaccinated (n = 6541) Not Vaccinated (n = 2928) tested for at a significant level except two (Influenza B Victoria
Sex and Influenza coinfections) (Table 5). Both Influenza B Victoria
Male 3916 (59.9) 1310 (44.7) and Influenza coinfections had reduced odds in the vaccinated
Female 2625 (40.1) 1618 (55.3)
cohort, but not at significant levels (Table 5). Examining non-
Beneficiary Category influenza viruses specifically, the odds of both coronavirus and
Active Duty 2516 (55.3) 247 (8.4)
human metapneumovirus in vaccinated individuals were signifi-
Child 1597 (24.4) 1469 (50.2)
Other 181 (2.8) 237 (8.1) cantly higher when compared to unvaccinated individuals
Retiree 403 (6.1) 272 (9.3) (OR = 1.36 and 1.51, respectively) (Table 5). Conversely, all other
Spouse 654 (10.0) 616 (21.0) non-influenza respiratory viruses had decreased odds in the vacci-
Unknown 90 (1.4) 87 (3.0)
nated population, including significantly decreased odds ratios in
Age Group vaccinated people with parainfluenza, RSV, and non-influenza
0–17 1616 (24.7) 1522 (52.0)
virus coinfections (Table 5). Additionally, the odds ratio in the no
18–35 3007 (46.0) 627 (21.4)
36+ 1918 (29.3) 779 (26.6)
pathogen detected cohort was significantly higher in vaccinated
versus unvaccinated individuals (OR = 1.51) (Table 5).
Disease Status
Other Respiratory Virus 2050 (31.3) 830 (28.3)
Influenza 2050 (31.3) 1299 (44.4)
No Pathogen Detected 2441 (37.4) 799 (27.3) 4. Discussion
Season
Winter 4327 (66.2) 1912 (65.3) Examining 6120 people with respiratory viruses other than
Spring 1644 (25.1) 548 (18.7) influenza and pan-negative results who submitted a respiratory
Summer 165 (2.5) 45 (1.5)
specimen for laboratory testing to the DoDGRS team, those who
Fall 405 (6.2) 423 (14.5)
received an influenza vaccine had a decreased risk of having other
respiratory pathogens identified compared to the unvaccinated
group. One study in the United States found similar results [12].
specimen resulted as influenza (Table 2). Winter was the predom- The study found influenza vaccination was not associated with
inate season for illness and specimen testing (Table 2). Among the detection of non-influenza respiratory viruses [12]. Additionally,
study population, 4549 people had AFCITA vaccination records the laboratory data in our study showed increased odds of coron-
(48.0%) while 4920 people self-reported vaccination status via avirus and human metapneumovirus in individuals receiving influ-
questionnaire (52.0%) (Data not shown). enza vaccination. The study finding similar results to our study
Examining the population with other respiratory viruses and no found no association between influenza vaccination and RSV, ade-
virus detected, those who were vaccinated had 19% lower unad- novirus, human metapneumovirus, rhinovirus or coronavirus [12].
justed odds (95% CI: 0.72, 0.91) of having other respiratory viruses The same study did find a significant association between parain-
compared to those who were unvaccinated (Table 3). After adjust- fluenza and influenza vaccination, but the association was in oppo-
ing for age group, the odds were 3% lower (95% CI: 0.86, 1.09) of site directions when comparing children and adults [12]. In our
having other respiratory viruses in the vaccinated population disease specific investigation, virus interference trends were
(Table 3). The unadjusted (data not shown) were statistically sig- noticed for coronavirus and human metapneumovirus; however,
nificant; however, the adjusted odds did not remain statistically two specific respiratory viruses (parainfluenza and RSV) showed
 G.G. Wolff / Vaccine 38 (2020) 350–354 353

Table 3
Virus interference odds ratio 2017–2018 influenza season.

Total Population
Other Respiratory Viruses Pan-Negative Respiratory Virus Unadjusted OR (95% CI) Adjusted OR (95% CI) Adjusted OR p-Value
Vaccinated 2050 2441 0.81 (0.72, 0.91) 0.97 (0.86, 1.09)* 0.60
Unvaccinated 830 799
AFCITA Confirmed Vaccination
Vaccinated 1417 2979 0.95 (0.68, 1.34) 1.23 (0.86, 1.76)** 0.25
Unvaccinated 51 102
Active Duty Only
Vaccinated 1046 2570 0.97 (0.73, 1.29) 1.20 (0.89, 1.61)** 0.24
Unvaccinated 73 174
*
Adjusted for age group.
**
Adjusted for age group and seasonality.

Table 4
Odds ratios for influenza cases versus controls using different control groups.

Cases vs All Controls

Case Control Unadjusted OR (95% CI) Adjusted OR (95% CI)*** Adjusted OR p-value
Vaccinated 2050 4491 0.57 (0.52, 0.63) 0.48 (0.43, 0.52) <0.0001
Unvaccinated 1299 1629
Cases vs Non-influenza Virus Positive Controls
Case Control Unadjusted OR (95% CI) Adjusted OR (95% CI)*** Adjusted OR p-value
Vaccinated 2050 2050 0.64 (0.57, 0.71) 0.51 (0.45, 0.57) <0.0001
Unvaccinated 1299 830
Cases vs Pan-Negative Controls
Case Control Unadjusted OR (95% CI) Adjusted OR (95% CI)*** Adjusted OR p-value
Vaccinated 2050 2441 0.52 (0.47, 0.57) 0.46 (0.41, 0.52) <0.0001
Unvaccinated 1299 799
***
Adjusted for gender, age group, and season.

Table 5
Respiratory viruses and odds ratios by vaccination status.

Virus Vaccinated (%) Not Vaccinated (%) OR (95% CI) P-Value

Influenza 2050 (31.3) 1299 (44.4) 0.57 (0.52, 0.63) <0.01
Influenza A 1256 (19.2) 741 (25.3) 0.70 (0.63, 0.78) <0.01
Influenza A H1N1 225 (3.4) 227 (7.8) 0.42 (0.35, 0.51) <0.01
Influenza A H3N2 1023 (15.6) 512 (17.5) 0.88 (0.78, 0.98) 0.02
Influenza B 662 (10.1) 474 (16.2) 0.58 (0.51, 0.66) <0.01
Influenza B Victoria 7 (0.1) 8 (0.3) 0.39 (0.14, 1.08) 0.07
Influenza B Yamagata 85 (1.3) 77 (2.6) 0.49 (0.36, 0.67) <0.01
Influenza Coinfection 9 (0.1) 9 (0.3) 0.45 (0.18, 1.13) 0.09
Non-Influenza Virus 2050 (31.3) 830 (28.3) 1.15 (1.05, 1.27) <0.01
Adenovirus 144 (2.2) 78 (2.7) 0.82 (0.62, 1.09) 0.17
Coronavirus 507 (7.8) 170 (5.8) 1.36 (1.14, 1.63) <0.01
Human Bocavirus 69 (1.1) 34 (1.2) 0.91 (0.60, 1.37) 0.64
Human Metapneumovirus 335 (5.1) 101 (3.5) 1.51 (1.20, 1.90) <0.01
No Pathogen Detected 2441 (37.3) 799 (27.3) 1.59 (1.44, 1.75) <0.01
Parainfluenza 139 (2.1) 92 (3.1) 0.67 (0.51, 0.87) <0.01
RSV 369 (5.6) 202 (6.9) 0.81 (0.68, 0.96) 0.02
Rhinovirus/Enterovirus 875 (13.4) 400 (13.7) 0.98 (0.86, 1.11) 0.71
Non-Influenza Virus Coinfection 225 (3.4) 138 (4.7) 0.72 (0.58, 0.89) <0.01

significant protection associated with influenza vaccine receipt, have influenza-like illness, collects a respiratory specimen, per-
and all others tested (adenovirus, human bocavirus, and rhi- forms diagnostic laboratory testing to determine the pathogen,
novirus/enterovirus) showed protection, although non-significant, and obtains the individual’s vaccination status [18–21]. The
associated with vaccination (Table 5). vaccine-associated virus interference phenomenon goes against
Additional examination of virus interference was accomplished the basic assumption of the test-negative vaccine effectiveness
by assessing the affect that three non-influenza control groups had study, that is, vaccination does not change the risk of infection with
on vaccine effectiveness (N = 9469). The adjusted ORs for the three other respiratory illness. The results of this study do not support a
groups ranged from 0.46 to 0.51, having similar 95% confidence potential for bias in test-negative influenza vaccine effectiveness
intervals, and accounting for a difference of 5% in vaccine effective- studies. In a test-negative study design, patients must be sick with
ness. The minute differences among the vaccine effectiveness of all influenza-like illness. Since the population must be ill, if the vacci-
three control groups does not support the virus interference nated population is more likely to have other respiratory viruses
concept. when compared to the non-vaccinated population, then in turn
Test-negative study designs are often utilized to calculate influ- they are less likely to have influenza. Bias introduced in these stud-
enza vaccine effectiveness. This type of study recruits subjects who ies may cause an overestimate of vaccine effectiveness.
354 G.G. Wolff / Vaccine 38 (2020) 350–354

Mandatory influenza vaccination is required for all Active Duty Declaration of Competing Interest
personnel [13]. As such, vaccination effectiveness studies examin-
ing strictly Active Duty military members have previously shown The authors declare that they have no known competing finan-
to be methodologically invalid and often times have uninter- cial interests or personal relationships that could have appeared
pretable results [22,23]. In order to examine potential issues with to influence the work reported in this paper.
mandatory vaccination, beneficiary category was included in the
logistic regression model. While beneficiary category did not References
remain significant, and was therefore not kept in the final logistic
model, Active Duty members were then separated from the rest [1] National Vaccine Information Center. What is the History of Influenza Vaccine
Use in America? <[Link]
of the population and virus interference among this population [Link]> 2019.
was calculated. Both the unadjusted and adjusted models did not [2] Treanor JJ, Talbot HK, Ohmit SE, et al. Effectiveness of seasonal influenza
show significant evidence of virus interference in Active Duty vaccines in the United States during a season with circulation of all three
vaccine strains. Clin Infect Dis 2012;55(7):951–9.
members; therefore, this large portion of the population does not [3] Ohmit SE, Thompson MG, Petrie JG, et al. Influenza vaccine effectiveness in the
appear to be skewing study results converse to the aforementioned 2011–2012 season: protection against each circulating virus and the effect of
vaccine effectiveness studies. prior vaccination on estimates. Clin Infect Dis 2014;58(3):319–27.
[4] McLean HQ, Thompson MG, Sundaram ME, et al. Influenza vaccine
Study limitations include the assumption of a causal relation-
effectiveness in the United States during 2012–2013: variable protection by
ship between influenza vaccination and respiratory viruses. Per- age and virus type. J Infect Dis 2015;211(10):1529–40.
haps there were other factors influencing rates of respiratory [5] Jackson ML, Chung JR, Jackson LA, et al. Influenza Vaccine Effectiveness in the
illnesses. Adjustment in the conditional logistic regression analysis United States during the 2015–2016 Season. N Engl J Med 2017;377
(6):534–43.
attempted to account for some factors which could influence respi- [6] Centers for Disease Control and Prevention. Seasonal Influenza Vaccine
ratory outcome. Additionally our study relied on self-reported vac- Effectiveness, 2004-2018. <[Link]
cination status when data were unavailable in AFCITA. Missing [Link]> 22 February 2019.
[7] Suzuki M, Camacho A, Ariyoshi K. Potential effect of virus interference on
vaccination status was especially high in non-Active Duty mem- influenza vaccine effectiveness estimates in test-negative designs. Epidemiol
bers of our study, as many of these patients were seen at clinics Infect 2014;142:2642–6.
off base, and therefore not necessarily tracked in AFCITA. However, [8] Cowling BJ, Fang VJ, Peiris JSM, et al. Increased risk of noninfluenza respiratory
virus infections associated with receipt of inactivated influenza vaccine. Clin
self-reported vaccination data were included to augment vaccina- Infect Dis 2012;54:1778–83.
tion status when AFCITA data were unavailable. Self-reported data [9] Feng S, Fowlkes A, Steffens A, Finelli L, Cowling B. Assessment of virus
were assumed to be relatively accurate. To ensure self-reported interference in a test-negative study of influenza vaccine effectiveness.
Epidemiology 2017;28(4):514–24.
data were not skewing findings, those with AFCITA records were [10] Cowling BJ, Nishiura H. Virus interference and estimates of influenza vaccine
examined exclusively for virus interference. Both the unadjusted effectiveness from test-negative studies. Epidemiology. 2012;23(6):930–1.
and adjusted models did not show significant evidence of virus [11] Rikin S, Jia H, Stockwell M, et al. Assessment of temporally-related acute
respiratory illness following influenza vaccination. Vaccine 2018;36:1958–64.
interference in AFCITA vaccinated personnel; therefore, vaccina-
[12] Sundaram ME, McClure DL, Belongia EA, et al. Influenza vaccination is not
tion record status does not appear to be skewing study results. associated with detection of noninfluenza respiratory viruses in seasonal
studies of influenza vaccine effectiveness. Clin Infect Dis 2013
September;57:789–93.
5. Conclusion [13] Defense Health Agency. Seasonal Influenza Vaccination Program (IVP). Falls
Church, Virginia: Defense Health Agency; 2018. p. 1–23. Interim Procedures
Virus interference associated with influenza vaccination has Memorandum 18-005.
[14] NxTAG Respiratory Pathogen Panel [package insert]. Toronto, Ontario, Canada:
been previously investigated [7–12]. However, this study was the Luminex; 2016.
first virus interference study conducted among highly vaccinated [15] Biofire FilmArray Respiratory Pathogen Panel [package insert]. Salt Lake City,
DoD personnel. The study included a diverse, well dispersed popu- Utah, USA; 28 June 2017.
[16] CDC Influenza A Subtyping Kit (Version 2) [package insert]. Atlanta, Georgia,
lation based on sex, age group, beneficiary category, and vaccina-
USA; 22 July 2016.
tion status. Additionally the population size was relatively large, [17] CDC Influenza A/B Subtyping Kit [package insert]. Atlanta, Georgia, USA; 23
and numerous respiratory pathogens were examined which were March 2016.
[18] Foppa I, Haber M, Ferdinands J, et al. The case test-negative design for studies
not previously individually investigated for virus interference.
of the effectiveness of influenza vaccine. Vaccine 2013;31(June):3104–9.
The overall results of the study showed little to no evidence [19] Jackson M, Nelson J. The test-negative design for estimating influenza vaccine
supporting the association of virus interference and influenza vac- effectiveness. Vaccine 2013;31(April):2165–8.
cination. Individual respiratory virus results were mixed, and some [20] Foppa I, Ferdinands J, Chaves S, et al. The case test-negative design for studies
of the effectiveness of influenza vaccine in inpatient settings. Int J Epidemiol
rebutted virus interference. Additionally those receiving the influ- 2016;6:2052.
enza vaccine were more likely to have no pathogen detected and [21] Sullivan S, Tchetgen Tchetgen E, Cowling B. Theoretical basis of the test-
reduced risk of influenza when compared to unvaccinated individ- negative study design for assessment of influenza vaccine effectiveness. Am J
Epidemiol 2016;5:345.
uals. Further research is necessary to help character virus interfer- [22] Coleman R, Eick-Cost A, Hawksworth AW, et al. Department of defense end-of-
ence and validate or refute the validity of the test-negative design season influenza vaccine effectiveness estimates for the 2017–2018 season.
for influenza vaccine effectiveness. MSMR 2018;25(10):16.
[23] Cooper M. DoD Influenza Surveillance and Vaccine Effectiveness. In: Oral
presentation at: The Vaccines and Related Biological Products Advisory
Disclaimer Committee (VRBPAC) Meeting; February 28, 2014; Silver Spring, MD.

The views expressed in this article are those of the authors and
do not necessarily represent the official policy or position of the
Department of Defense, or the U.S. Government.