PDF Archive

Easily share your PDF documents with your contacts, on the Web and Social Networks.

Share a file Manage my documents Convert Recover PDF Search Help Contact



siriwardena2014 .pdf


Original filename: siriwardena2014.pdf
Title: Influenza and pneumococcal vaccination and risk of stroke or transient ischaemic attack—Matched case control study
Author: A. Niroshan Siriwardena

This PDF 1.7 document has been generated by Elsevier / Acrobat Distiller 9.0.0 (Windows), and has been sent on pdf-archive.com on 18/02/2017 at 15:37, from IP address 89.78.x.x. The current document download page has been viewed 245 times.
File size: 576 KB (8 pages).
Privacy: public file




Download original PDF file









Document preview


Vaccine 32 (2014) 1354–1361

Contents lists available at ScienceDirect

Vaccine
journal homepage: www.elsevier.com/locate/vaccine

Influenza and pneumococcal vaccination and risk of stroke
or transient ischaemic attack—Matched case control study
A. Niroshan Siriwardena a,∗ , Zahid Asghar b , Carol C.A. Coupland c
a
Community and Health Research Unit, University of Lincoln & Lincolnshire Community Health Services NHS Trust, School of Health and Social Care,
College of Social Science, Brayford Campus, Lincoln LN6 7TS, UK
b
Community and Health Research Unit, University of Lincoln, School of Health and Social Care, College of Social Science, Brayford Campus,
Lincoln LN6 7TS, UK
c
University of Nottingham, Division of Primary Care, School of Community Health Sciences, Floor 13, Tower Building, Nottingham NG7 2RD, UK

a r t i c l e

i n f o

Article history:
Received 26 October 2013
Received in revised form
17 December 2013
Accepted 14 January 2014
Available online 28 January 2014
Keywords:
Influenza vaccination
Pneumococcal vaccination
Influenza
Stroke
Transient ischemic attack
Matched case-control study

a b s t r a c t
Background: Evidence that respiratory infections trigger stroke suggests that influenza or pneumococcal
vaccination might prevent stroke. We aimed to investigate whether influenza or pneumococcal vaccination or both together were associated with reduced risk of stroke or transient ischaemic attack
(TIA).
Methods: We used a matched 1:1 case-control design with data from the United Kingdom General Practice
Research Database. Cases, aged 18 years or above with stroke (fatal or non-fatal) and TIA during September
2001 to August 2009, were compared with controls matched for age, sex, calendar time and practice,
adjusting for cardiovascular risk factors, vaccine risk groups, comorbidity and indicators of functional
ability.
Results: We included 26,784 cases of stroke and 20,227 cases of TIA with equal numbers of matched
controls. Influenza vaccination within-season was associated with 24% reduction in stroke risk (adjusted
OR 0.76, 95% CI 0.72 to 0.80) but no reduction in TIA (1.03, 0.98 to 1.09). Stroke risk was significantly
lower with early (September to mid-November: 0.74, 0.70 to 0.78) but not later influenza vaccination
(mid-November onwards: 0.92, 0.83 to 1.01). Associations persisted after multiple imputation of missing
data and sensitivity analysis for unmeasured confounders. Pneumococcal vaccination was not associated
with a reduction in risk of stroke (0.98, 0.94 to 1.00) or TIA (1.15, 1.08 to 1.23).
Conclusions: Influenza vaccination was associated with a 24% reduction in risk of stroke but not TIA.
Pneumococcal vaccination was not associated with reduced risk of stroke or TIA. This has important
implications for potential benefits of influenza vaccine.
© 2014 Elsevier Ltd. All rights reserved.

1. Introduction
Stroke is an important cause of death, disability and long term
illness worldwide. It affects 150,000 people in the UK annually
causing substantial social and economic effects [1,2].
Classical risk factors, such as increasing age, hypertension,
smoking, diabetes and heart disease,[3] account for only 50–60%
of strokes [4] raising the possibility of other important triggers [5].
Stroke is more common in winter [6]. This may be due to seasonal
changes in metabolic risk factors [7] or cold-induced vascular stress
[8] but both may also be linked to respiratory (e.g. influenza and

∗ Corresponding author. Tel.: +44 1522 886939; fax: +44 1522 837058.
E-mail addresses: nsiriwardena@lincoln.ac.uk (A.N. Siriwardena),
zasghar@lincoln.ac.uk (Z. Asghar), carol.coupland@nottingham.ac.uk
(C.C.A. Coupland).
0264-410X/$ – see front matter © 2014 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.vaccine.2014.01.029

pneumococcal infections) since stroke mortality mirrors peaks in
winter influenza and pneumonia [9]. This poses the question of
whether influenza or pneumococcal vaccine could prevent a proportion of the remaining cases.
Several case-control studies have shown increased likelihood of
respiratory symptoms one to four weeks before strokes occur [5],
implying that early treatment or prevention of respiratory infection
might also prevent stroke. A similar link has been shown between
respiratory infection and myocardial infarction [10]. Antibiotics
have been found to be ineffective [11] and antivirals of doubtful
effectiveness [12], but influenza or pneumococcal vaccination may
be preventative for strokes.
Some observational studies have found influenza vaccination to
be associated with a reduced risk of stroke [13–17], either alone
or combined with pneumococcal vaccine [18], but other studies
have not [19–21]. Because of this inconclusive evidence, we aimed

A.N. Siriwardena et al. / Vaccine 32 (2014) 1354–1361

to investigate the association between influenza or pneumococcal
vaccination and stroke.
2. Methods
2.1. Study design and setting
We used a matched nested case-control study design. Data were
extracted from the General Practice Research Database (GPRD),
now called the Clinical Practice Research Datalink (CPRD), a large
computerized anonymised database representative of and comprising over 5% of the population of England and Wales [22].
In England and Wales there are over 8000 general practices, each
with an average of around 7000 patients cared for by 4.3 whole time
equivalent general practitioners and their ancillary staff (including
nurses, healthcare assistants and administrative staff) providing
primary care [23]. Each person has a unique identifier (the NHS
number), and all influenza and pneumococcal vaccinations are provided free to those patients in risk groups.
Influenza vaccination in the UK is recommended for people aged
65 years and over and people aged less than 65 years with at-risk
conditions including chronic respiratory, heart, renal, liver or neurological disease, diabetes, immunosuppression, pregnancy, long
stay residential or nursing care, and carers of elderly or disabled
people [24].
Vaccines are administered and recorded by practices and where
these are (rarely) administered elsewhere (e.g. the workplace)
practices are informed so they can be entered on practice records.
Vaccine recording in the GPRD has been validated against national
data from the UK vaccination monitoring programme [25]. The
average influenza vaccination uptake in England in 2011/2012 was
74% for patients aged 65 years and above and around 51.6% for those
patients aged less than 65 years who were in a risk group [26].
Coding for vaccination, stroke, diseases, health behaviours,
administrative and outcome data use a standard UK hierarchical
(Read or Multilex) coding system developed for general practice
with each disease code linked to ICD codes. Previous GPRD validation studies found that medical data including demographic
information, health behaviours, specialist referrals, hospital admissions and treatment outcomes, with high quality, validated clinical
information including stroke, deaths and dates of incident (or index
date) recording [27–29].
2.2. Selection of cases and controls
The study cohort was drawn from all GPRD practices over eight
years from 1/09/2001 to 31/08/2009. Cases were patients aged 18
years and over, recorded with standard computer codes for stroke
or transient ischemic attack (TIA).
Each case of stroke or TIA was matched to one control according
to age, sex, general practice attended and calendar time. Controls
were patients registered at the same practice during the study
period identified at the same index date as the corresponding case
to account for possible seasonal effects and effects due to duration of observation for events. Controls were selected at random
(and before their exposure status was known to reduce selection
bias) from the pool of eligible matched controls for each case using
incidence density sampling according to person-time at risk [30].
Controls had to be alive and not transferred out of the practice or
dead prior to the index date of their matched case.
All cases and controls that had less than five years of clinical
records before the index date on the database were excluded to
ensure completeness of recording of exposures and confounding
variables. Those with a previous diagnosis of stroke or TIA were
also excluded.

1355

2.3. Outcome and exposure measures
Outcome measures were unadjusted and adjusted odds ratios
and 95% confidence intervals for stroke or TIA associated with vaccination prior to the index date. Since we used incident cases and
matched controls to cases by calendar time, from a dynamic population, odds ratios were taken as estimates of rate ratios [30].
We used the recorded vaccination date before the index date.
Where patients had more than one vaccination recorded in the
same season (1st September to 31st August of the following year),
only the first recorded date was used since duplicate entries are
usually errors or repeated for administrative purposes.
For the main analyses, influenza vaccination was defined as
vaccination given in the same season preceding the index date
(1st September to end of August of the following year before
the index date). Other influenza vaccination exposures considered
were: early (1st September to 15th November) or late vaccination
(16th November to 28th/29th February). Time since last influenza
vaccination at the index date was defined as 0–3, 3–6, 6–12, or more
than 12 months before the index date.
Pneumococcal vaccination was defined as vaccination at any
time before the index date because in the UK the recommendation is for a once in a lifetime dose of vaccine for those in most risk
groups. Combined vaccination was defined as pneumococcal vaccine ever and influenza vaccine within season. Patients who had not
received combined vaccination were classified as having received
influenza vaccination only, pneumococcal vaccination only or neither of the vaccinations.
2.4. Confounders and effect modifiers
We adjusted for known confounding variables, in particular vaccine risk groups, cardiovascular risk factors, treatments and other
factors (Charlson comorbidity index [31], GP consultations and
home visits) recorded with a computer (Read or OXMIS) code [see
Table 1].
We accounted for prevention bias and the ‘healthy vaccinee’
effect by taking other preventive and lifestyle factors into account
such as weight, smoking, cholesterol level, and cardiovascular
treatments including statins. Functional status is also an important
potential confounder [32,33]. Although data for functional status
were not available, GP consultation rate is a measure of healthcare utilisation which is related to functional and economic status
[34–36] and repeated home visits are more likely to be required by
people who are more frail [37], so we also accounted for these in
this study.
In order to further account for functional status and unknown
confounders we conducted a sensitivity analysis to ascertain the
potential effect of an unknown or unmeasured confounder on the
estimated odds ratio [38].
2.5. Data analysis
Data were analysed using STATA, version 11.0. Descriptive
statistics were given in terms of frequencies for categorical variables and means with standard deviations or medians with
interquartile ranges for continuous variables depending on the distribution.
We used conditional logistic regression for matched case control studies, calculating unadjusted and adjusted odds ratios with
95% confidence intervals for stroke or TIA according to influenza
vaccination (within season), and pneumococcal vaccination (ever)
in cases compared with controls. Analyses were also carried out for
time since influenza vaccination and for early and late in-season
vaccinations. Adjusted analyses accounted for the confounding
variables listed. We also adjusted the analyses of pneumococcal

1356

A.N. Siriwardena et al. / Vaccine 32 (2014) 1354–1361

Table 1
Characteristics of cases of stroke and TIA combined and their matched controls.
Variables

Cases N = 47,011, N (%)

Controls N = 47,011, N (%)

Unadjusted odds ratio (95% CI)

Matching variables
Aged <65 years
Aged ≥65 years

10,910 (23.2)
36,101 (76.8)

10,910 (23.2)
36,101 (76.8)

NA*
NA*

22,584 (48.0)
24,427 (52.0)

22,584 (48.0)
24,427 (52.0)

NA*

11,713 (24.9)
7631 (16.2)
5767 (12.3)
4121 (8.7)
178 (0.4)
298 (0.6)
143 (0.3)
20 (0.1)

8084 (17.2)
6844 (14.6)
4114 (8.8)
3415 (7.3)
90 (0.2)
184 (0.4)
76 (0.2)
9 (<0.1)

1.64
1.14
1.47
1.28
1.98
1.61
1.88
2.22

1.59 to 1.70
1.10 to 1.18
1.40 to 1.54
1.21 to 1.34
1.53 to 2.55
1.35 to 1.95
1.42 to 2.49
1.01 to 4.88

22,952 (48,8)
1668 (3.6)
6191 (13.2)
4216 (9.0)
8044 (17.1)

18,332 (39.0)
1046 (2.2)
5164 (11.0)
3569 (7.6)
7206 (15.3)

1.54
1.64
1.25
1.21
1.15

1.50 to 1.58
1.51 to 1.77
1.20 to 1.31
1.16 to 1.27
1.10 to 1.20

19,707 (41.9)
22,680 (48.2)
3944 (8.4)
680 (1.5)
19,009 (42.3)
27.8 [4.9]
16,832 (37.5)
164.5 [26.3]
95.2 [12.9]
24,411 (54.0)
5.4 [2.0]

21,854 (46.5)
20,148 (42.9)
3439 (7.3)
1570 (3.3)
17,446 (39.2)
27.4 [4.8]
14,844(34.0)
160.6 [25.8]
92.7 [12.3]
20,715 (45.9)
5.3 [1.9]

1.00
1.28
1.31



1.25 to 1.33
1.24 to 1.38


1.01†

1.00 to 1.02

1.00†
1.02†

1.00 to 1.01
1.01 to 1.02

1.00‡

0.99 to 1.02

5095 (10.8)
7007 (14.9)
7683 (16.3)
1304 (2.8)

2143 (4.6)
5289 (11.3)
3796 (8.1)
628 (1.3)

2.55
1.38
2.25
2.12

2.42 to2.69
1.33 to 1.44
2.16 to 2.35
1.92 to 2.33

1 [0 to 3]

0 [0 to 2]

15,573 (34.5)
8603 (19.1)
6669 (14.8)
4470 (9.9)
3065 (6.8)
6760 (15.0)

24,476 (54.2)
6774 (15.0)
5376 (11.9)
2817 (6.2)
2040 (4.5)
3657 (8.1)

1.00
2.09
2.08
2.72
2.60
3.19

2.01 to 2.17
1.99 to 2.17
2.57 to 2.86
2.44 to 2.76
3.04 to 3.35

8.7 [8.1]
26 [12 to 45]

7.3 [7.1]
20 [8 to 38]

12,239 (26.0)
11,647 (24.8)
11,062 (23.5)
12,063 (25.7)
1 to 4 (2)

16,177 (34.4)
12,197 (25.9)
9674 (20.6)
8963 (19.1)
1 to 4 (2)

1.00
1.31
1.62
2.03

1.26 to 1.36
1.56 to 1.69
1.95 to 2.11

43,788 (93.1)
2795 (6.0)
428 (0.9)

44,822 (95.3)
1883 (4.0)
306 (0.7)

1.00
1.60
1.58

1.50 to 1.71
1.36 to 1.84

Sex
Male
Female
Vaccine risk groups
Chronic heart disease
Asthma/COPD
Diabetes
Chronic renal disease
Chronic liver disease
Chronic neurological disease
Splenectomy
Immunosuppression or HIV
Cardiovascular risk factors
Hypertension
Peripheral vascular disease
Hyperlipidaemia
Family history of stroke
Family history of ischaemic heart disease (IHD)
Smoking status (latest recorded)
Never smoked
Ex-smoker
Current smoker
Not recorded
Body mass index recorded within past 5 years**
Body mass index kg/m2 (mean [SD])
Systolic blood pressure recorded in past 3 years
Systolic blood pressure mmHg (mean [SD])
Diastolic blood pressure mmHg (mean [SD])
Total cholesterol recorded
Total cholesterol mmol/l (mean [SD])
Treatments
Aspirin uptake (≥2 prescriptions in previous 6 months)
Antihypertensive treatment (≥1 prescription in previous year)
Statin uptake (≥1 prescription in previous year)
Oral anticoagulant (≥1 prescription in previous year)
Comorbidity index
Charlson index: median [IQR]
Charlson index
0
1
2
3
4
≥5
GP consultations
GP surgery consultation rate (visits per year): mean [SD]
GP consultations: median [IQR] **
GP surgery consultations**
≤12
13–26
27–44
≥45
GP home visits IQR (median)**
GP home visits**
≤1
2–10
≥11
*
**



Not applicable because we matched for these variables.
In the past 5 years.
Per 10 unit increase.
Per unit increase.

vaccination for influenza vaccination and vice versa. Analyses were
repeated for patients aged 65 years and over since this is a target
group for vaccination, and for those aged under 65 years comparing the results with a test for interaction. We also calculated odds
ratios for influenza vaccination in the previous year for separate categories of month of index date (September to November, December
to March, April to August).

Multiple imputation was carried out to replace missing values
for smoking status, systolic blood pressure, body mass index and
total cholesterol using the ICE command in Stata. Ten imputed
datasets were created and the results were combined using Rubin’s
rules to account for uncertainty in the imputed data [39]. This analysis adjusted for systolic blood pressure, body mass index and total
cholesterol as well as the confounders adjusted for previously.

A.N. Siriwardena et al. / Vaccine 32 (2014) 1354–1361

An identical analysis was conducted for stroke and TIA separately.
2.6. Ethical approval
We received ethical.approval for the study from the Independent Scientific Advisory Committee (ISAC) of the GPRD, the National
Research Ethics Service (NRES) through existing approval of observational studies using GPRD data and NHS Lincolnshire Research
Management and Governance (previously Lincolnshire Teaching
Primary Care Trust) [05/MRE04/87]. The study also received ethical approval from the University of Lincoln, School of Health and
Social Sciences Ethics Committee.

1357

Influenza vaccination in adults is recommended for people aged
65 years and over and people aged less than 65 years in risk groups.
Because of the overlap between vaccine risk groups and risk of
stroke, vaccinated persons, whether influenza or pneumococcal
vaccine, were more likely to be in a vaccine risk group, except
for chronic liver disease, chronic neurological disease, splenectomy
or immunosuppression/HIV, and also more likely to be taking cardiovascular drugs than unvaccinated people (Table 2). Vaccinated
individuals were also more likely to have cardiovascular risk factors except, current smoking (Table 2), and more comorbidities.
Finally, vaccinated patients consulted more frequently overall and
had more home visits.
3.1. Influenza vaccination

3. Results
We included 47,011 cases comprising 26,784 cases of stroke and
20,227 cases of TIA with equal numbers of matched controls in
the study (Table 1). Cases were significantly more likely to be in
a vaccine risk group, have cardiovascular risk factors (except high
BMI, blood pressure or total cholesterol), higher levels of comorbidity and greater numbers of GP consultations and home visits than
controls (Table 1).

Influenza vaccination given within the same season (as the index
date) was associated with a significant 24% reduction in the risk of
stroke (adjusted OR 0.76, 95% CI 0.72 to 0.80) after adjusting for confounding variables (Table 3). Comorbidity (based on the Charlson
index) and number of GP consultations in the five years before the
index date were the main confounding variables. Stroke risk was
significantly lower with early (September to mid-November: 0.74,
0.70 to 0.78) but not later influenza vaccination (mid-November

Table 2
Distribution of risk factors by vaccination status.
Risk factor

Vaccine risk groups
Chronic heart disease
Asthma/COPD
Diabetes
Chronic renal disease
Chronic liver disease
Chronic neurological disease
Splenectomy
Immunosuppression
Cardiovascular risk factors
Hypertension
Peripheral vascular disease
Hyperlipidaemia
Family history of stroke
Family history of IHD
Smoking status (latest recorded):
Never smoked
Ex-smoker
Current smoker
Not recorded
Obese or overweight
Treatments
Aspirin treatment
Antihypertensive treatment
Statin treatment
Oral anticoagulants
Consultations
GP home visits** :
≤1
2–10
≥11
GP surgery visits** :
≤12
13–26
27–44
≥45
Comorbidity index:
0
1
2
3
4
≥5
**

In the past 5 years.

No vaccination
N = 26,175

Influenza vaccination only
within the previous year
N = 7021

Pneumococcal vaccination
only administered ever
N = 12,153

Combined vaccination
N = 48,673

3476 (9.4)
2886 (7.8)
1558 (4.2)
1008 (2.7)
77 (0.2)
181 (0.5)
14 (<0.1)
5 (<0.1)

2463 (18.2)
1698 (12.5)
1073 (7.9)
516 (3.8)
27 (0.2)
73 (0.5)
8 (0.1)
4 (<0.1)

2879 (32.3)
2076 (23.4)
1426 (16.0)
1133 (12.7)
44 (0.5)
65 (0.7)
61 (0.7)
5 (0.1)

8701 (28.3)
6452 (21.0)
4835 (15.7)
3846 (12.5)
90 (0.3)
139 (0.5)
114 (0.4)
14 (0.1)

9705 (26.2)
578 (1.6)
2438 (6.6)
2286 (6.2)
4557 (12.3)

6095 (44.9)
346 (2.6)
1646 (12.1)
1189 (8.8)
2179 (16.0)

4492 (50.3)
430 (4.8)
1193 (13.4)
713 (8.0)
1348 (15.2)

16,636 (54.2)
1113 (3.6)
5271 (17.2)
2887 (9.4)
5618 (18.2)

14,644 (39.5)
12,285 (33.2)
4030 (10.9)
6101 (16.5)
20,173 (54.5)

6211 (45.8)
6418 (47.3)
819 (6.0)
116 (0.9)
5804 (42.7)

4018 (45.0)
4082 (45.7)
643 (7.2)
185 (2.1)
3400 (38.3)

12,964 (42.2)
16,428 (53.5)
1262 (4.1)
74 (0.2)
11,513 (37.4)

1470 (4.0)
3167 (8.6)
2513 (6.8)
331 (0.9)

1221 (9.0)
2462 (18.1)
1971 (14.5)
311 (0.9)

381 (4.3)
690 (7.8)
745 (8.4)
171 (2.3)

3052 (9.9)
5427 (17.6)
5729 (18.6)
1022 (1.9)

36,003 (97.2)
934 (2.5)
123 (0.3)

13,163 (97)
346 (2.6)
55 (0.4)

7629 (85.5)
1102 (12.3)
197 (2.2)

28,360 (92.3)
2052 (6.7)
316 (1)

19,145 (51.7)
8883 (24.0)
5293 (14.3)
3739 (10.1)

2354 (17.4)
4107 (30.3)
3676 (27.1)
3427 (25.3)

2602 (29.1)
2395 (26.8)
1974 (22.1)
1957 (21.9)

3248 (10.6)
7505 (24.4)
8970 (29.2)
11,005 (35.8)

23,155 (62.5)
5542 (15)
3559 (9.6)
1829 (4.9)
1115 (3.0)
1860 (5.0)

6163 (45.4)
2886 (21.3)
1986 (14.6)
1015 (7.5)
566 (4.2)
948 (7.0)

1805 (20.2)
1480 (16.6)
1494 (16.7)
1089 (12.2)
871 (9.8)
2189 (24.5)

8926 (29.1)
5469 (17.8)
5006 (16.3)
3354 (10.9)
2553 (8.3)
5420 (17.6)

1358

A.N. Siriwardena et al. / Vaccine 32 (2014) 1354–1361

Table 3
Association between influenza vaccination and stroke.
Cases N (%)
N = 26,784

Within season vaccination
13,547 (50.6)
Same season as index date
11,883 (44.4)
Early within-season (Sept. to
mid-Nov.)
1664 (6.2)
Late within-season
(mid-Nov. to Feb.)
Within season vaccination by age
1200 (19.2)
<65 years
≥65 years
12,347 (60.2)
Early within-season (Sept. to mid-Nov.)
<65 years
972 (15.6)
≥65 years
10,911 (53.2)
Late within-season (mid-Nov. to Feb.)
228 (3.6)
<65 years
1436 (7.0)
≥65 years
Time since last vaccination at index date
7774 (29.0)
Never vaccinated
4495 (16.8)
0–3 months
3–6 months
4119 (15.4)
7355 (27.5)
6–12 months
>12months
3041 (11.4)
Vaccination in previous year by month of index date†
September–November
3456 (53.1)
5387 (59.2)
December–March
April–August
6805 (60.9)

Unadjusted

Adjusted*

OR

95% CI

OR

95% CI

OR

95% CI

13,605 (50.8)
12,160 (45.4)

0.99
0.97

0.95 to 1.03
0.93 to 1.01

0.76
0.74

0.72 to 0.80
0.70 to 0.78

0.81
0.79

0.77 to 0.85
0.74 to 0.85

1443 (5.4)

1.16

1.07 to 1.26

0.92

0.83 to 1.01

0.98

0.89 to 1.07

806 (12.9)
12,797 (62.4)

1.67
0.88

1.50 to 1.84
0.84 to 0.92

0.80
0.74

0.69 to 0.92
0.70 to 0.78

0.87
0.80

0.76 to 0.99
0.76 to 0.84

663 (10.6)
11,497 (56.0)

1.64
0.87

1.47 to 1.83
0.83 to 0.91

0.79
0.73

0.67 to 0.92
0.69 to 0.77

0.87
0.78

0.75 to 1.00
0.74 to 0.82

143 (2.3)
1300 (6.3)

1.80
1.02

1.44 to 2.23
0.94 to 1.12

0.84
0.89

0.63 to 1.13
0.80 to 0.98

0.86
0.95

0.66 to 1.12
0.85 to 1.04

8774 (32.8)
4713 (17.6)
3933 (14.7)
7079 (26.4)
2285 (8.5)

1.00
1.07
1.30
1.28
1.58

1.00 to 1.15
1.21 to 1.39
1.21 to 1.36
1.48 to 1.69

1.00
0.78
0.89
0.89
1.22

0.72 to 0.85
0.82 to 0.98
0.82 to 0.95
1.12 to 1.31

0.78
0.89
0.89
1.22

0.72 to 0.85
0.82 to 0.98
0.82 to 0.95
1.12 to 1.31

3475 (53.4)
5376 (59.1)
6627(59.3)

0.99
1.00
1.09

0.92 to 1.06
0.94 to 1.08
1.03 to 1.16

0.77
0.75
0.84

0.69 to 0.85
0.69 to 0.82
0.78 to 0.91

0.80
0.82
0.93

0.73 to 0.89
0.75 to 0.89
0.86 to 1.01

Controls N (%)
N = 26,784

Multiply imputed
adjusted**

*
Adjusted for chronic heart disease, asthma/COPD, diabetes, chronic renal failure, chronic liver disease, splenectomy, immunosuppression/HIV, Charlson (comorbidity)
index, hypertension, peripheral vascular disease, hyperlipidaemia, smoking status, family history of stroke/TIA, family history of AMI, aspirin uptake, antihypertensive
treatment, statin uptake, number of home visits and general practice consultations.
**
Adjusted for all of the above and for BMI, total cholesterol and systolic blood pressure using multiple imputed data.

The totals for each of the periods were: 11,181 cases and 11,181 controls for the April to August period; 6504 cases and 6504 controls for September–November period;
and 9099 cases and 9099 controls for the December–March period.

onwards: 0.92, 0.83 to 1.01) [Wald’s test comparing odds ratios
p < 0.001]. Influenza vaccination was associated with a reduced risk
of stroke within, but not after, 12 months from vaccination and for
index dates from September to March but not beyond this consistent with the timing of circulating influenza virus.
The association of influenza vaccination was marginally greater
for those aged 65 years and over (adjusted OR 0.74, 95% CI 0.70 to
0.78) compared to those aged under 65 years (adjusted OR 0.80, 95%
CI 0.69 to 0.92) but the difference was not statistically significant
(test for interaction: p = 0.32). Associations persisted after multiple
imputation of missing data.
Influenza vaccination within-season was not associated with a
reduction in risk of TIA (1.03, 0.98 to 1.09) (Table 4).
3.2. Sensitivity analyses
We explored the potential effect of an unmeasured confounder
such as frailty using sensitivity analysis (Appendices A and B: Table
7). This showed that, assuming the prevalence of an unmeasured
confounder such as functional ability to be present in 20% of the
population, and to be a risk factor for stroke (with an odds ratio of 2),
and that subjects with the unmeasured confounder were 30% less
likely to be vaccinated (OR for vaccination 0.7), then the odds ratio
for the association between influenza vaccination and stroke would
increase by 0.06 to OR 0.80 (95% CI 0.75–0.84) and would remain
statistically significant. The association remained significant even
for an unmeasured confounder being present in 40% and having a
strong association with stroke and vaccination status.
3.3. Pneumococcal vaccination
After adjusting for confounding variables, including influenza
vaccination, there was no significant association between

pneumococcal vaccination and risk of stroke (0.98, 0.94 to 1.00)
or TIA (1.15, 1.08 to 1.23) (Tables 5 and 6).
4. Discussion
4.1. Main findings
Influenza vaccination within-season was associated with 24%
reduction in stroke risk (adjusted OR 0.76, 95% CI 0.72 to 0.80)
but no reduction in TIA (1.03, 0.98 to 1.09). The risk of stroke was
significantly lower with early (September to mid-November: 0.74,
0.70 to 0.78) but not later influenza vaccination (mid-November
onwards: 0.92, 0.83 to 1.01). Associations persisted after multiple
imputation of missing data and sensitivity analysis for unmeasured
confounders. Pneumococcal vaccination was not associated with a
reduction in risk of stroke (0.98, 0.94 to 1.00) or TIA (1.15, 1.08 to
1.23).
The reduction in risk of stroke occurred within six months of
influenza vaccination and for strokes occurring between September
and March. Comorbidity and GP consultation rate in the five years
before the index date were the main confounding variables.
4.2. Strengths and limitations
The matched case-control design was efficient although there
are known problems of bias and confounding [40]. The large sample
gave ample power to detect associations in a representative population. We reduced selection bias by including all cases of stroke
within the selected time period [41] and eligible randomly selected
controls, free of the outcome of interest, matched to cases independent of the exposure of interest [42]. Analysis matched for age,
sex, practice and calendar time reduced confounding from these
sources and increased precision.

A.N. Siriwardena et al. / Vaccine 32 (2014) 1354–1361

1359

Table 4
Association between influenza vaccination and transient ischaemic attack.
Unadjusted

Adjusted*

Multiply imputed
adjusted**

OR

95% CI

OR

95% CI

OR*

95% CI

10,266 (50.8)
9154 (45.3)
1112 (5.5)

1.32
1.31
1.35

1.26 to 1.38
1.25 to 1.38
1.23 to 1.48

1.03
1.02
1.13

0.98 to 1.09
0.96 to 1.08
1.02 to 1.25

1.04
1.02
1.14

0.98 to 1.09
0.96 to 1.09
1.02 to 1.26

663 (14.3)
9603 (61.6)

1.81
1.22

1.62 to 2.03
1.16 to 1.29

0.97
1.02

0.84 to 1.13
0.96 to 1.08

0.98
1.03

0.84 to 1.13
0.97 to 1.09

519 (11.2)
8635 (55.4)

1.88
1.22

1.67 to 2.14
1.15 to 1.28

1.00
1.01

0.85 to 1.17
0.95 to 1.07

1.00
1.01

0.85 to 1.18
0.95 to 1.08

144 (3.1)
968 (6.2)

1.58
1.26

1.26 to 1.97
1.13 to 1.39

0.88
1.13

0.67 to 1.16
1.01 to 1.26

0.89
1.14

0.68 to 1.16
1.02 to 1.27

6494 (32.1)
3547 (17.5)
3082 (15.2)
5321 (26.3)
1783 (8.8)

1.00
1.45
1.50
1.52
1.40

1.34 to 1.57
1.38 to 1.63
1.41 to 1.62
1.29 to 1.51

1.00
1.08
1.11
1.09
1.14

1.00 to 1.18
1.01 to 1.21
1.01 to 1.18
1.04 to 1.23

1.00
1.10
1.13
1.11
1.15

1.00 to 1.21
1.03 to 1.23
1.02 to 1.20
1.06 to 1.26

2776 (54.8)
4062 (59.3)
4939 (59.4)

1.24
1.36
1.35

1.14 to 1.36
1.26 to 1.47
1.26 to 1.45

0.99
1.12
0.99

0.89 to 1.10
1.03 to 1.23
0.91 to 1.08

1.00
1.12
1.01

0.89 to 1.11
1.01 to 1.24
0.92 to 1.11

Controls N (%)
N = 20,227

Cases N (%)
N = 20,227

Within season vaccination
11,150 (55.1)
Same season as index date
9905 (49.0)
Early within-season (Sept. to mid-Nov.)
Late within-season (mid-Nov. to Feb.)
1245 (6.2)
Within season vaccination by age
1032 (22.2)
<65 years
10,118 (64.9)
≥65 years
Early within-season (Sept. to mid-Nov.)
833 (17.9)
<65 years
≥65 years
9072 (58.2)
Late within-season (mid-Nov. to Feb.)
<65 years
199 (4.3)
1046 (6.7)
≥65 years
Time since last vaccination at index date
5326 (26.3)
Never vaccinated
3825 (18.9)
0–3 months
3359 (16.6)
3–6 months
5827 (28.8)
6–12 months
>12 months
1890 (9.3)
Vaccination in previous year by month of index date†
September–November
2987 (59.0)
4435 (64.8)
December–March
5369 (64.6)
April–August

*
Adjusted for chronic heart disease, asthma/COPD, diabetes, chronic renal failure, chronic liver disease, splenectomy, immunosuppression/HIV, Charlson (comorbidity)
index, hypertension, peripheral vascular disease, hyperlipidaemia, smoking status, family history of stroke/TIA, family history of AMI, aspirin uptake, antihypertensive
treatment, statin uptake, number of home visits and general practice consultations.
**
Adjusted for all of the above and for BMI, total cholesterol and systolic blood pressure using multiple imputed data.

The totals for each of the periods were: 8314 cases and 8314 controls for the April to August period; 5066 cases and 5066 controls for September–November period; and
6847 cases and 6847 controls for the December–March period.

Table 5
Association between pneumococcal vaccination and stroke.
Cases
N (%)

Unadjusted

Controls
N (%)

Pneumococcal vaccine ever received
All ages
17,206 (64.2)
16,773 (62.6)
2371 (37.9)
1379 (22.0)
<65 years
14,835 (72.3)
15,394(75.0)
≥65 years

Adjusted (but not
for influenza
vaccination)*

Adjusted (including
for influenza
vaccination)*

Multiply imputed
adjusted (including for
influenza vaccination)**

OR*

95% CI

OR

95% CI

OR

95% CI

OR

95% CI

1.10
2.49
0.86

1.05 to 1.14
2.29 to 2.74
0.82 to 0.90

0.86
0.82
0.86

0.81 to 0.90
0.69 to 0.98
0.81 to 0.90

0.94
0.91
0.96

0.89 to 1.00
0.75 to 1.10
0.90 to 1.02

0.96
0.88
1.00

0.92 to 1.02
0.74 to 1.05
0.94 to 1.05

*
Adjusted for chronic heart disease, asthma/COPD, diabetes, chronic renal failure, chronic liver disease, splenectomy, immunosuppression/HIV, Charlson (comorbidity) index,
hypertension, peripheral vascular disease, hyperlipidaemia, smoking status, family history of stroke/TIA, family history of AMI, aspirin uptake, antihypertensive treatment,
statin uptake, number of home visits and general practice consultations.
**
Adjusted for all of the above and for BMI, total cholesterol and systolic blood pressure using multiple imputed data.

Potential sources of bias include misclassification and missing
data. To reduce misclassification, patients were only selected if they
had at least five years of recorded data. Although we observed missing values for smoking status, systolic blood pressure, body mass

index and total cholesterol we used standard multiple imputation
procedures to account for these and obtained similar results to
our complete case analysis. Since only data recorded prior to the
outcomes of interest were included and because neither cases nor

Table 6
Association between pneumococcal vaccination and transient ischaemic attack.
Cases
N (%)

Controls
N (%)

Pneumococcal vaccine ever received
14,100 (69.7)
12,747 (63.0)
All ages
1877 (40.4)
1123 (24.2)
<65 years
12,223 (78.4)
11,624 (74.6)
≥65 years

Unadjusted

Adjusted (but not
for influenza
vaccination)*

Adjusted (including
for influenza
vaccination)*

Multiply imputed
adjusted (including for
influenza vaccination)**

OR

95% CI

OR

95% CI

OR

95% CI

OR

95% CI

1.51
2.63
1.27

1.44 to 1.59
2.34 to 2.92
1.20 to 1.35

1.14
1.54
1.03

1.08 to 1.20
1.36 to 1.77
0.97 to 1.09

1.15
1.61
1.02

1.08 to 1.23
1.40 to 1.84
0.95 to 1.10

1.16
1.61
1.03

1.09 to 1.23
1.40 to 1.85
0.96 to 1.10

*
Adjusted for chronic heart disease, asthma/COPD, diabetes, chronic renal failure, chronic liver disease, splenectomy, immunosuppression/HIV, Charlson (comorbidity)
index, hypertension, peripheral vascular disease, hyperlipidaemia, smoking status, family history of stroke/TIA, family history of AMI, aspirin uptake, antihypertensive
treatment, statin uptake, number of home visits and general practice consultations.
**
Adjusted for all of the above and for BMI, total cholesterol and systolic blood pressure using multiple imputed data.

1360

A.N. Siriwardena et al. / Vaccine 32 (2014) 1354–1361

controls were asked to remember information, particularly vaccination status, prior to the event date the effect of recall bias which
may have been a limitation in some previous studies [13,16] was
removed.
We did not adjust our analyses for changes in vaccine strain or
efficacy: for example the 2009 H1N1 influenza pandemic, which
was associated with cardiovascular as well as respiratory deaths,
occurred at the end of our study inclusion period [43].
The limitations of the case-control approach include confounding and residual unmeasured confounding, particularly from
‘healthy vaccinee’ bias, where healthier people are more likely to
be vaccinated, or ‘indication bias’, where those at higher risk are
more likely to be offered vaccination.
Indication bias is likely to have led to the reversal in the
unadjusted and adjusted odds ratios for the association between
influenza vaccination and stroke in patients aged less than 65
years old. Adjusted and unadjusted odds ratios for the association
between influenza vaccination and stroke both showed a reduction
of risk of stroke (albeit greater in the adjusted odds ratios) in those
aged over 65 years because uptake is recommended for everyone
in this age group and therefore less affected by indication.
The large validated database we used enabled us to adjust
for important confounders including clinical risk groups for vaccination, cardiovascular risk factors and differences in treatment
between cases and controls.
We also adjusted for additional comorbidity and attempted to
account for functional ability and frailty using GP consultation and
home visit rates, all of which may have been, and were in fact,
associated with the probability of receiving vaccination and/or risk
of stroke. Information about physical activity and diet were not
available and these factors may be a source of residual confounding.
The lack of association of influenza vaccination with TIA, the lack
of association with pneumococcal vaccination with both TIA and
stroke, and the stability of the association between influenza and
reduction in stroke risk to sensitivity analysis increases the validity
of our positive findings for reduction in stroke risk associated with
influenza vaccination. Even an unmeasured confounder with a very
strong association with both stroke and vaccination status would
not have eliminated the association we found between influenza
vaccination and stroke.
The biological plausibility of a within-season effect of influenza
vaccination preventing influenza related stroke, given that
influenza vaccination is only likely to be effective against circulating strains of influenza virus, strengthens our findings.
If the positive protective effect for influenza vaccination within
the previous 12 months was due to confounding by indication,
we would also expect, but did not find, a positive effect beyond
12 months. We believe this also supports the plausibility of our
findings and that the difference between associations for early and
late vaccination and stroke cannot be explained by confounding by
indication.

4.3. Comparison with previous studies
This study confirms a beneficial effect of influenza vaccination
for preventing stroke but did not find any benefit for pneumococcal
vaccination in preventing stroke. Our findings are consistent with
our previous studies showing an association between influenza
vaccination and reduction in risk of AMI, particularly for early
influenza vaccination [36,44], but no reduction in risk with
pneumococcal vaccination [36]. A recent systematic review of
randomised control trials, also found that influenza vaccine was
associated with a lower risk of major adverse cardiovascular events
[45]. Our findings support growing evidence that better prevention
of influenza may prevent some cardiovascular events.

4.4. Implications for policy and practice
Our findings reinforce current recommendations for annual
influenza vaccination [46] with the potential added benefit for
stroke prevention. This may help improve suboptimal vaccination
rates, particularly those groups of patients who might benefit. The
results also suggest that early influenza vaccination is associated
with greater reduction in stroke risk compared with later vaccination which encourages practices to complete the vaccination
programme early.
The mechanism by which respiratory infections or influenza
might precipitate stroke is not known. Various suggestions have
been postulated including non-specific immune stimulation [47];
fever leading to endothelial dysfunction, hypercoagulability or
increased viscosity; and stress or changes in metabolic risk factors
in response to infection. Influenza may cause atheroma [48], affect
carotid plaque stability or promote rupture [49].
There remains uncertainty about whether to extend vaccination to younger adults at risk of stroke. A self-controlled case
series study may reduce further the effect of unknown or unmeasured confounders [44] but experimental studies are needed before
recommending changes in indications or timing of vaccination.

4.5. Conclusion
The finding of an association between reduced risk of stroke
with influenza vaccination, but not pneumococcal vaccination,
supports the current influenza vaccination programme with the
added potential benefit of stroke prevention. This could result in
improvements in patient and population health, if the benefit is
confirmed by experimental studies and leads to higher vaccination rates. Randomised controlled studies are needed to investigate
the effect of influenza vaccination on primary prevention of stroke
outcomes.

Conflict of interest statement
All authors declare that no support from any organisation for the
submitted work; no financial relationships with any organisations
that might have an interest in the submitted work in the previous
three years; no other relationships or activities that could appear
to have influenced the submitted work.

Contributors
A.N.S. and C.C. had the initial idea for the study and had the
main responsibility for the conception and design of the study. Z.A.
undertook the data analysis supported by C.C. A.N.S. wrote the first
draft of the paper. All authors interpreted the results, participated
in reviewing and editing the entire report, and approved the final
version to be published.

Acknowledgements
We are grateful to the Clinical Practice Research Datalink (CPRD)
for providing the data for the study. This work presents independent research supported by the National Institute for Health
Research (NIHR) under its Research for Patient Benefit (RfPB) Programme (Grant Reference Number PB-PG-0808-16254). The views
expressed are those of the authors and not necessarily those of the
NHS, the NIHR or the Department of Health.

A.N. Siriwardena et al. / Vaccine 32 (2014) 1354–1361

Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/j.vaccine.
2014.01.029.
References
[1] Giles MF, Rothwell PM. Substantial underestimation of the need for outpatient
services for TIA and minor stroke. Age Ageing 2007;36(6):676–80.
[2] Wolfe CD. The impact of stroke. Br Med Bull 2000;56(2):275–86.
[3] Carroll K, Murad S, Eliahoo J, Majeed A. Stroke incidence and risk factors in a
population-based cohort study. London: Office of National Statistics; 2001.
[4] Ware JH. The limitations of risk factors as prognostic tools. N Engl J Med
2006;355(25):2615–7.
[5] Emsley HC, Hopkins SJ. Acute ischaemic stroke and infection: recent and emerging concepts. Lancet Neurol 2008;7(4):341–53.
[6] Kelly-Hayes M, Wolf PA, Kase CS, Brand FN, McGuirk JM, D’Agostino RB. Temporal patterns of stroke onset. The Framingham Study. Stroke 1995;26(8):1343–7.
[7] Woodhouse PR, Khaw KT, Plummer M, Foley A, Meade TW. Seasonal variations
of plasma fibrinogen and factor VII activity in the elderly: winter infections and
death from cardiovascular disease. Lancet 1994;343(8895):435–9.
[8] The Eurowinter Group. Cold exposure and winter mortality from ischaemic
heart disease, cerebrovascular disease, respiratory disease, and all causes in
warm and cold regions of Europe. Lancet 1997;349(9062):1341–6.
[9] Field TS, Zhu H, Tarrant M, Mitchell JR, Hill MD. Relationship between supraannual trends in influenza rates and stroke occurrence. Neuroepidemiology
2004;23(5):228–35.
[10] Warren-Gash C, Smeeth L, Hayward AC. Influenza as a trigger for acute myocardial infarction or death from cardiovascular disease: a systematic review.
Lancet Infect Dis 2009;9(10):601–10.
[11] Brassard P, Bourgault C, Brophy J, Kezouh A, Suissa S. Antibiotics in primary
prevention of stroke in the elderly. Stroke 2003;34(9):e163–6.
[12] Cooper NJ, Sutton AJ, Abrams KR, Wailoo A, Turner D, Nicholson KG. Effectiveness of neuraminidase inhibitors in treatment and prevention of influenza A
and B: systematic review and meta-analyses of randomised controlled trials.
BMJ 2003;326(7401):1235.
[13] Lavallee P, Perchaud V, Gautier-Bertrand M, Grabli D, Amarenco P. Association between influenza vaccination and reduced risk of brain infarction. Stroke
2002;33(2):513–8.
[14] Nichol KL, Nordin J, Mullooly J, Lask R, Fillbrandt K, Iwane M. Influenza vaccination and reduction in hospitalizations for cardiac disease and stroke among
the elderly. N Engl J Med 2003;348(14):1322–32.
[15] Smeeth L, Thomas SL, Hall AJ, Hubbard R, Farrington P, Vallance P. Risk of
myocardial infarction and stroke after acute infection or vaccination. N Engl
J Med 2004;351(25):2611–8.
[16] Grau AJ, Fischer B, Barth C, Ling P, Lichy C, Buggle F. Influenza vaccination is
associated with a reduced risk of stroke. Stroke 2005;36(7):1501–6.
[17] Vila-Corcoles A, Ochoa-Gondar O, Rodriguez-Blanco T, Gutierrez-Perez A, VilaRovira A, Gomez F, et al. Clinical effectiveness of pneumococcal vaccination
against acute myocardial infarction and stroke in people over 60 years: the
CAPAMIS study, one-year follow-up. BMC Public Health 2012;12:222.
[18] Hung IF, Leung AY, Chu DW, Leung D, Cheung T, Chan CK, et al. Prevention of
acute myocardial infarction and stroke among elderly persons by dual pneumococcal and influenza vaccination: a prospective cohort study. Clin Infect Dis
2010;51(9):1007–16.
[19] Pinol-Ripoll G, de lP I, Santos S, Purroy F, Mostacero E. Chronic bronchitis
and acute infections as new risk factors for ischemic stroke and the lack of
protection offered by the influenza vaccination. Cerebrovasc Dis 2008;26(4):
339–47.
[20] Clayton TC, Thompson M, Meade TW. Recent respiratory infection and risk of
cardiovascular disease: case-control study through a general practice database.
Eur Heart J 2008;29(1):96–103.
[21] Tseng HF, Slezak JM, Quinn VP, Sy LS, Van den Eeden SK, Jacobsen SJ. Pneumococcal vaccination and risk of acute myocardial infarction and stroke in men.
JAMA 2010;303(17):1699–706.
[22] Walley T, Mantgani A. The UK general practice research database. Lancet
1997;350(9084):1097–9.
[23] NHS Information Centre for Health and Social Care. General and personal medical services NHS Staff 2002–2012. London: Health and Social Care Information
Centre; 2013.

1361

[24] Department of Health. Seasonal flu plan: Winter 2012/13. London: HMSO;
2012.
[25] Joseph C, Goddard N, Gelb D. Influenza vaccine uptake and distribution in
England and Wales using data from the General Practice Research Database,
1989/90–2003/04. J Public Health (Oxf) 2005;27(4):371–7.
[26] Department of Health, From the Chief Medical Officer. The flu immunisation
programme 2012/13. London: HMSO; 2012.
[27] Jick SS, Kaye JA, Vasilakis-Scaramozza C, Garcia Rodriguez LA, Ruigomez A,
Meier CR, et al. Validity of the general practice research database. Pharmacotherapy 2003;23(5):686–9.
[28] Herrett E, Thomas SL, Schoonen WM, Smeeth L, Hall AJ. Validation and validity
of diagnoses in the General Practice Research Database: a systematic review.
Br J Clin Pharmacol 2010;69(1):4–14.
[29] Khan NF, Harrison SE, Rose PW. Validity of diagnostic coding within the
General Practice Research Database: a systematic review. Br J Gen Pract
2010;60(572):e128–36.
[30] Knol MJ, Vandenbroucke JP, Scott P, Egger M. What do case-control studies estimate? Survey of methods and assumptions in published case-control research.
Am J Epidemiol 2008;168(9):1073–81.
[31] Charlson ME, Pompei P, Ales KL, Mackenzie CR. A new method of classifying
prognostic comorbidity in longitudinal studies: development and validation. J
Chronic Dis 1987;40(5):373–83.
[32] Jackson LA, Jackson ML, Nelson JC, Neuzil KM, Weiss NS. Evidence of bias
in estimates of influenza vaccine effectiveness in seniors. Int J Epidemiol
2006;35(2):337–44.
[33] Jackson LA, Nelson JC, Benson P, Neuzil KM, Reid RJ, Psaty BM, et al. Functional
status is a confounder of the association of influenza vaccine and risk of all
cause mortality in seniors. Int J Epidemiol 2006;35(2):345–52.
[34] Carr-Hill RA, Rice N, Roland M. Socioeconomic determinants of rates of consultation in general practice based on fourth national morbidity survey of general
practices. BMJ 1996;312(7037):1008–12.
[35] Jordan K, Ong BN, Croft P. Previous consultation and self reported health status
as predictors of future demand for primary care. J Epidemiol Community Health
2003;57(2):109–13.
[36] Siriwardena AN, Gwini SM, Coupland CA. Influenza vaccination, pneumococcal vaccination and risk of acute myocardial infarction: matched case-control
study. CMAJ 2010;182(15):1617–23.
[37] Lone NI, Simpson C, Kavanagh K, Robertson C, McMenamin J, Ritchie L, et al.
Seasonal influenza vaccine effectiveness in the community (SIVE): protocol
for a cohort study exploiting a unique national linked data set. BMJ Open
2012;2(2):e001019.
[38] Groenwold RH, Nelson DB, Nichol KL, Hoes AW, Hak E. Sensitivity analyses to
estimate the potential impact of unmeasured confounding in causal research.
Int J Epidemiol 2010;39(1):107–17.
[39] Rubin DB. Multiple imputation for nonresponse in surveys. Hoboken, NJ: WileyInterscience; 2004.
[40] Delgado-Rodriguez M, Llorca J. Bias. Epidemiol Community Health
2004;58(8):635–41.
[41] Schulz KF, Grimes DA. Case-control studies: research in reverse. Lancet
2002;359(9304):431–4.
[42] Grimes DA, Schulz KF. Compared to what? Finding controls for case-control
studies. Lancet 2005;365(9468):1429–33.
[43] Dawood FS, Iuliano AD, Reed C, Meltzer MI, Shay DK, Cheng PY, et al. Estimated global mortality associated with the first 12 months of 2009 pandemic
influenza A H1N1 virus circulation: a modelling study. Lancet Infect Dis
2012;12(9):687–95.
[44] Gwini SM, Coupland CA, Siriwardena AN. The effect of influenza vaccination on
risk of acute myocardial infarction: self-controlled case-series study. Vaccine
2011;29(6):1145–9.
[45] Udell JA, Zawi R, Bhatt DL, Keshtkar-Jahromi M, Gaughran F, Phrommintikul A,
et al. Association between influenza vaccination and cardiovascular outcomes
in high-risk patients: a meta-analysis. JAMA 2013;310(16):1711–20.
[46] Fiore AE, Shay DK, Broder K, Iskander JK, Uyeki TM, Mootrey G, et al. Prevention and control of influenza: recommendations of the Advisory Committee on
Immunization Practices (ACIP). MMWR Recomm Rep 2008;57(RR-7):1–60.
[47] Lehr HA, Sagban TA, Ihling C, Zahringer U, Hungerer KD, Blumrich M, et al.
Immunopathogenesis of atherosclerosis: endotoxin accelerates atherosclerosis
in rabbits on hypercholesterolemic diet. Circulation 2001;104(8):914–20.
[48] Ravnskov U, McCully KS. Infections may be causal in the pathogenesis of
atherosclerosis. Am J Med Sci 2012;344(5):391–4.
[49] Keller TT, van der Meer JJ, Teeling P, van der SK, Idu MM, Rimmelzwaan GF,
et al. Selective expansion of influenza A virus-specific T cells in symptomatic
human carotid artery atherosclerotic plaques. Stroke 2008;39(1):174–9.


Related documents


siriwardena2014
tripedialb
global pneumococcal vaccine market
united states pediatric vaccines market
age of autism vaccination outcomes
mmillerpdf


Related keywords