Effectiveness of Influenza Vaccination on Hospitalizations and Risk Factors for Severe Outcomes in Hospitalized Patients With COPD

[

Original Research COPD

]

Effectiveness of Influenza Vaccination on Hospitalizations and Risk Factors for Severe Outcomes in Hospitalized Patients With COPD Sunita Mulpuru, MD; Li Li, MSc; Lingyun Ye, MSc; Todd Hatchette, MD; Melissa K. Andrew, MD, PhD; Ardith Ambrose, RN; Guy Boivin, MD; William Bowie, MD; Ayman Chit, MBiotech; Gael Dos Santos, PhD; May ElSherif, MD; Karen Green, MSc; Francois Haguinet, MSc; Scott A. Halperin, MD; Barbara Ibarguchi, MSc; Jennie Johnstone, MD; Kevin Katz, MD; Joanne M. Langley, MD; Jason LeBlanc, PhD; Mark Loeb, MD; Donna MacKinnon-Cameron, MMath; Anne McCarthy, MD; Janet E. McElhaney, MD; Allison McGeer, MD; Jeff Powis, MD; David Richardson, MD; Makeda Semret, MD; Vivek Shinde, MD, MPH; Daniel Smyth, MD; Sylvie Trottier, MD; Louis Valiquette, MD; Duncan Webster, MD; Shelly A. McNeil, MD, FIDSA; on behalf of the Serious Outcomes Surveillance (SOS) Network of the Canadian Immunization Research Network (CIRN)*

The effectiveness of influenza vaccination in reducing influenza-related hospitalizations among patients with COPD is not well described, and influenza vaccination uptake remains suboptimal.

BACKGROUND:

Data were analyzed from a national, prospective, multicenter cohort study including patients with COPD, hospitalized with any acute respiratory illness or exacerbation between 2011 and 2015. All patients underwent nasopharyngeal swab screening with polymerase chain reaction (PCR) testing for influenza. The primary outcome was an influenzarelated hospitalization. We identified influenza-positive cases and negative control subjects and used multivariable logistic regression with a standard test-negative design to estimate the vaccine effectiveness for preventing influenza-related hospitalizations. METHODS:

Among 4,755 hospitalized patients with COPD, 4,198 (88.3%) patients with known vaccination status were analyzed. The adjusted analysis showed a 38% reduction in influenzarelated hospitalizations in vaccinated vs unvaccinated individuals. Influenza-positive patients (n ¼ 1,833 [38.5%]) experienced higher crude mortality (9.7% vs 7.9%; P ¼ .047) and critical illness (17.2% vs 12.1%; P < .001) compared with influenza-negative patients. Risk factors for mortality in influenza-positive patients included age > 75 years (OR, 3.7 [95% CI, 0.4-30.3]), cardiac comorbidity (OR, 2.0 [95% CI, 1.3-3.2]), residence in long-term care (OR, 2.6 [95% CI, 1.5-4.5]), and home oxygen use (OR, 2.9 [95% CI, 1.6-5.1]).

RESULTS:

CONCLUSIONS: Influenza vaccination significantly reduced influenza-related hospitalization among patients with COPD. Initiatives to increase vaccination uptake and early use of antiviral agents among patients with COPD could reduce influenza-related hospitalization and critical illness and improve health-care costs in this vulnerable population. TRIAL REGISTRY:

ClinicalTrials.govNo.:NCT01517191; URL www.clinicaltrials.gov CHEST 2019; 155(1):69-78

KEY WORDS:

chestjournal.org

COPD; hospitalization; influenza; pulmonary; vaccine effectiveness

69

COPD exacerbations (COPDE) are a leading cause of hospital admissions and re-admissions in North America.1 COPD care costs the health system more than half a billion Canadian dollars annually, largely due to hospital-related care. Respiratory viral infections commonly trigger COPDE, accounting for 40% to 60% of infectious exacerbations. Previous studies reported a 5% to 22% prevalence of influenza virus among patients experiencing an exacerbation.2-8 International guidelines recommend annual influenza vaccination for patients with COPD to reduce influenzarelated exacerbations and acute respiratory infections.9,10 This recommendation is based on a systematic review of six randomized studies that showed a reduction in exacerbations following influenza vaccination, particularly in the first 4 weeks postvaccination.11 Two studies in this review evaluated the effectiveness of influenza vaccination on hospitalizations among outpatients with COPD;

Patients and Methods A post hoc analysis of data collected in a national multicenter prospective cohort study was conducted to evaluate the effectiveness of influenza vaccination in adult patients hospitalized with acute respiratory illness.17,18 Data from the Canadian Immunization Research Network Serious Outcomes Surveillance (SOS) database were analyzed. The SOS database contains information from adult patients ($ 16 years of age) hospitalized for acute respiratory infection (including pneumonia, acute COPDE or asthma, unexplained sepsis, other respiratory infections, or cases with respiratory or influenza-like illness symptoms). Data were collected during influenza seasons between 2011 and 2015. The influenza season was defined according to Canadian influenza surveillance data as November 1 to April 30.17-19

ABBREVIATIONS:

COPDE = COPD exacerbations; NP = nasopharyngeal; PCR = polymerase chain reaction; SOS = Serious Outcomes Surveillance AFFILIATIONS: From the Ottawa Hospital Research Institute (Drs Mulpuru and McCarthy), Clinical Epidemiology Program, University of Ottawa, ON, Canada; Canadian Center for Vaccinology (Messrs Li and Ye; Drs Hatchette, Andrew, ElSherif, Halperin Langley, LeBlanc, and McNeil; and Mss Ambrose and MacKinnon-Cameron), IWK Health Centre and Nova Scotia Health Authority, Dalhousie University, Halifax, NS, Canada; Centre Hospitalier Universitaire de Québec (Drs Boivin and Trottier), Quebec, QC, Canada; Faculty of Medicine, Division of Infectious Diseases, Department of Medicine (Dr Bowie), University of British Columbia, Vancouver, BC, Canada; Sanofi Pasteur (Mr Chit), Swiftwater, PA; Leslie Dan Faculty of Pharmacy (Mr Chit), University of Toronto, Toronto, ON, Canada; Business & Decision Life Sciences (Dr Dos Santos), Brussels, Belgium (GlaxoSmithKline); Mount Sinai Hospital (Ms Green and Dr McGeer), Toronto, ON, Canada; GlaxoSmithKline (Mr Haguinet), Wavre, Belgium; GlaxoSmithKline (Ms Ibarguchi), Mississauga, ON, Canada; Division of Infectious Diseases, Department of Medicine (Dr Johnstone) and Division of Infectious Diseases, Department of Pathology and Molecular Medicine (Dr Loeb), McMaster University, Hamilton, ON, Canada; North York General Hospital (Dr Katz), Toronto, ON, Canada; Health Sciences North Research Institute (Dr McElhaney), Sudbury, ON, Canada; Toronto East General Hospital (Dr Powis), Toronto, ON,

70 Original Research

however, these studies were limited by small sample sizes, a single season, and single-center evaluations.12,13 Despite evidence-based recommendations, only 50% to 60% of patients with COPD are vaccinated.14,15 Furthermore, specific influenza-related outcomes in the COPD population have not been directly evaluated, and the inference of poor outcomes in this population comes from studies of elderly patients with a mix of chronic lung diseases.16 Understanding the severity of influenzarelated complications, risk factors for poor outcomes, and influenza vaccine effectiveness to prevent hospitalizations may improve knowledge translation and policy initiatives to increase compliance with immunization guidelines in this population. The aim of the present study was to determine the effectiveness of the influenza vaccine in preventing influenza-related hospitalizations and to describe the severity of influenza-related complications and risk factors for poor outcomes among hospitalized patients with COPD.

Forty-six hospitals across five Canadian provinces participated in the study.17,18 Hospitalized adults with a diagnosis of COPD documented on the medical record were included. Study monitors screened admissions to the ICUs and medical wards to identify and obtain consent from eligible patients. All study participants received influenza testing by nasopharyngeal (NP) swab with polymerase chain reaction (PCR), as described in the original study protocol. Patient demographic characteristics, medical history, medications, influenza vaccination status, symptoms, hospital course, length of stay, and complications during hospitalization were collected from the medical record and by patient interview. Only standard-dose trivalent inactivated influenza vaccines were administered during the study period. Canada; William Osler Health System (Dr Richardson), Brampton, ON, Canada; Department of Medicine, Division of Infectious Diseases, Faculty of Medicine (Dr Semret), McGill University, Montreal, QC, Canada; GlaxoSmithKline (Dr Shinde), King of Prussia, PA; The Moncton Hospital (Dr Smyth), Moncton, NB, Canada; Université de Sherbrooke (Dr Valiquette), Sherbrooke, QC, Canada; and Horizon Health (Dr Webster), Saint John, NB, Canada. Dr Dos Santos is currently at GlaxoSmithKline (Wavre, Belgium); Ms Ibarguchi is currently at Bayer Inc. (Mississauga, ON, Canada); and Dr Shinde is currently at Novavax Vaccines (Washington, DC). *Collaborators from the Serious Outcomes Surveillance (SOS) Network of the Canadian Immunization Research Network (CIRN) are listed in the Acknowledgements. FUNDING/SUPPORT: Multiple funding sources facilitated and sustained the SOS Network, including the Canadian Institutes for Health Research [FRN#96974], the Public Health Agency of Canada, and a collaborative Research Grant from GlaxoSmithKline Biological SA. CORRESPONDENCE TO: Sunita Mulpuru, MD, The Ottawa Hospital, 501 Smyth Rd, Box 211, Ottawa, Ontario, K1H 8L6, Canada; e-mail: [email protected] Copyright Ó 2018 American College of Chest Physicians. Published by Elsevier Inc. All rights reserved. DOI:

https://doi.org/10.1016/j.chest.2018.10.044

[

155#1 CHEST JANUARY 2019

]

The primary outcome was an influenza-related hospitalization (NP swab positive for influenza). Secondary outcomes included ICU admission, all-cause mortality up to 30 days’ postdischarge, use of noninvasive positive pressure ventilation, mechanical ventilation, and length of stay. Clinical outcomes were ascertained by study personnel through contact with the patient or patient designate. If data were not available at day 30, patients who were not receiving palliative care, who were improving at discharge, and who were not readmitted to the same hospital were considered to have survived. A test-negative design was used to determine vaccine effectiveness for influenza-related hospitalizations and included all patients with known influenza vaccination status. The test-negative design is a modified casecontrol method that is validated among inpatient populations with acute respiratory infection.20,21 In this study, patients with known vaccination status seeking health care for influenza-like illness were identified, uniformly tested for influenza, and grouped into influenza-positive cases and influenza-negative control subjects.20,22 A multivariate regression model was used to determine the association between influenza vaccination and hospitalization. Vaccine effectiveness was estimated by: (1 – OR of influenza-related hospitalization) in vaccinated vs unvaccinated individuals, multiplied by 100.

disease, diabetes, and cancer), dwelling (long-term care or assisted living), smoking status, baseline oxygen use, BMI, previous physician or ED visit, and number of medications prior to admission (# 4 or > 4 medications). Regression models to determine vaccine effectiveness were adequately powered to potentially include > 100 independent variables without introducing significant bias.23 Among patients with laboratory-confirmed influenza, baseline demographic characteristics and outcomes were described by using means  SD, medians (interquartile ranges), and proportions. Oneway ANOVA and c2 testing was used to compare continuous and categorical variables between COPD patients with and without home oxygen use. Multivariable logistic regression was used to determine risk factors for mortality, ICU care, use of noninvasive positive pressure ventilation, and mechanical ventilation among patients with confirmed influenza. Independent predictor variables were selected a priori for each model and tested with univariate analyses. Backward selection was used, ensuring that clinically relevant and confounding variables remained.

The multivariable regression model to determine the association between influenza vaccination and hospitalization in this study was adjusted for age categories (16-49, 50-64, 65-75, and > 75 years), influenza season, comorbidities (heart disease, liver disease, renal

The study protocol was approved by the Research Ethics Boards of participating institutions, including Dalhousie University, Nova Scotia, Canada (approval no. 1019858). The study was registered in the United States National Institutes of Health database.24

Results

Tables 1 and 2 describe the baseline characteristics, vaccine status, and clinical outcomes in this cohort.

A total of 4,755 adult patients with COPD were identified who were hospitalized with acute respiratory infection or exacerbation during the study. PCRconfirmed influenza was diagnosed in 38.5% (1,833 of 4,755) of patients. Influenza Vaccine Effectiveness

Overall, 4,198 of 4,755 (88.3%) hospitalized patients with COPD and known vaccination status were included in the influenza vaccine effectiveness analysis (Fig 1).

Compared with hospitalized patients without influenza infection, patients with influenza (1,490 of 4,198 [35.5%]) were older (age > 75 years; 50.8% vs 47.6%), more likely to be current smokers (34.4% vs 27.2%), to reside in long-term care (9.2% vs 7.0%), and less likely to be vaccinated during the season of hospitalization (58.9% vs 70.6%). Influenzainfected patients more frequently required mechanical ventilation (8.7% vs 5.2%) and experienced higher crude mortality (9.7% vs 7.9%) and need for critical care

Patients with COPD, hospitalized during influenza seasons 2011-2015 N = 4,755

Influenza Positive n = 1,833

Vaccinated n = 878

Unvaccinated n = 612

Influenza Negative n = 2,922

Unknown Status n = 343

Vaccinated n = 1,912

Unvaccinated n = 796

Unknown Status n = 214

Figure 1 – Patients with COPD who were hospitalized for acute respiratory illness or exacerbation during winter seasons, 2011 to 2015.

chestjournal.org

71

TABLE 1

] Risk Factors for Influenza-Related Hospitalizations Among Adult Patients With COPD and Known Vaccination Status Who Were Hospitalized Between 2011 and 2015 Influenza Seasons Influenza Positive (n ¼ 1490)

Influenza Negative (n ¼ 2708)

16-49 y

41 (2.8%)

57 (2.1%)

50-64 y

309 (20.7%)

513 (18.9%)

Variable Age categories

P Value .001

65-75 y

383 (25.7)

848 (31.3%)

> 75 y

757 (50.8%)

1,290 (47.6%)

Female sex

771 (51.7%)

1,360 (50.2%)

.35

23 (1.5%)

32 (1.2%)

.31

501 (34.4%)

725 (27.2%)

< .001

Heart disease

769 (51.6%)

1,528 (56.4%)

Diabetes

381 (25.6%)

710 (26.2%)

Renal or liver disease

249 (16.7%)

469 (17.3%)

.62

Cancer

297 (19.9%)

578 (21.3%)

.28

Aboriginal ethnicitya Current smokerb Comorbidities

Solid organ transplant Rheumatic disease Long-term care resident

.003 .83

9 (0.6%)

15 (0.6%)

.84

67 (4.5%)

99 (3.7%)

.18

137 (9.2%)

189 (7.0%)

.01

Assisted living

172 (11.5%)

230 (8.5%)

Oxygen use at homec

151 (10.1%)

556 (20.5%)

< .01

.001

Prior physician or ED visit for current illness

490 (32.9%)

669 (24.7%)

< .01

Influenza vaccinationd

878 (58.9%)

1,912 (70.6%)

< .01

1,203 (80.7%)

2,320 (85.7%)

< .001

>4 Medications at admission a

Aboriginal ethnicity not known in 42 of 1,490 case subjects and in 53 of 2,708 control subjects. Smoking status unknown in 32 of 1,490 case subjects and 40 of 2,708 control subjects. c Oxygen use at home unknown in 1 of 1,490 case subjects, and 1 of 2,708 control subjects. d Indicates trivalent influenza vaccine given in influenza season during which patient was admitted to the hospital. b

(17.2% vs 12.1%) compared with influenza-negative patients. Using the test-negative design, the seasonal influenza vaccine reduced influenza-related hospitalizations among patients with COPD by 37.5% (95% CI, 27.346.2) compared with unvaccinated patients (Table 3). The crude (unadjusted) estimate for influenza vaccine effectiveness was 40.3% (95% CI, 31.8-47.7). The predominant influenza strains and corresponding vaccine effectiveness are reported for each study year in Table 3. Vaccine effectiveness remained between 43% and 49% during the first 3 years of the study and was only 6% during the final year of the study due to a vaccine/influenza strain mismatch. Outcomes Among Patients With Confirmed Influenza and COPD

Table 4 describes the baseline characteristics and outcomes of influenza-positive patients. Approximately one half were aged > 75 years (51.5%), and one third

72 Original Research

were active smokers (33.2%). One fifth (22% [407 of 1,833]) of patients were hospitalized from a long-term care or assisted living facility, and 10.2% (187 of 1,833) were oxygen dependent. Among patients with influenza and known vaccination status, 41% (612 of 1,490) had not received the influenza vaccination during the season when hospitalization occurred, and 69.3% of patients received an antiviral medication in the hospital. Approximately 1 of 10 influenza-positive patients required mechanical ventilation (9.7% [177 of 1,833]), and 11.1% (204 of 1,833) died. One of five patients was admitted to the ICU (18.1% [332 of 1,833]), and the mean  SD length of stay was 11  13.3 days. Risk Factors for Poor Outcomes Among InfluenzaPositive Patients

Tables 5 and 6 report risk factors for 30-day mortality and ICU admission among hospitalized patients with COPD. Home oxygen use was significantly associated with higher mortality and ICU admission. The crude mortality rate was significantly higher in the vaccinated

[

155#1 CHEST JANUARY 2019

]

TABLE 2

] Outcomes for Influenza-Related Hospitalizations Among Adult Patients With COPD and Known Vaccination Status Who Were Hospitalized Between the 2011 and 2015 Influenza Seasons Influenza Positive (n ¼ 1,490)

Variable

Influenza Negative (n ¼ 2,708)

Total (N ¼ 4,198)

30-d mortality

145 (9.7%)

215 (7.9%)

360 (8.6%)

ICU admission

257 (17.2%)

329 (12.1%)

586 (14.0%)

Use of NIV during hospitalization

159 (10.7%)

366 (13.5%)

525 (12.5%)

130 (8.7%)

142 (5.2%)

272 (6.5%)

7 (5-13)

7 (5-11)

7 (5-12)

Mechanical ventilation Length of hospital stay, median (IQR), d IQR ¼ interquartile range; NIV ¼ noninvasive ventilation.

group vs the unvaccinated group (11.0% vs 7.8%; P < .001). However, in the adjusted analysis, vaccination was not associated with a significant reduction in mortality (OR, 0.9 [95% CI, 0.6-1.4]). Risk factors for mechanical ventilation included diabetes (OR, 1.7 [95% CI, 1.1-2.6]) and current smoking (OR, 3.2 [95% CI, 1.5-7.2]). Risk factors for use of noninvasive ventilation included home oxygen use (OR, 2.0 [95% CI, 1.3-3.2]). Home oxygen use was the most significant factor associated with longest length of stay in hospital (OR, 3.5 [95% CI, 1.5-3.7]).

in influenza-related hospitalizations (38%). Our estimates of seasonal vaccine effectiveness against hospitalization in this study mirror those reported by the Centers for Disease Control and Prevention in the general outpatient population. The available vaccine was not an accurate match for the predominant circulating strain for H3N2 during the 2014 to 2015 season, which accounted for the low vaccine effectiveness estimate in that year.25,26

Discussion

Influenza burden was high (38.5%) among hospitalized patients, and influenza-infected patients experienced significant crude mortality (1 of 10) and critical illness (1 of 5). Despite guideline recommendations, only 66.5% of patients were vaccinated in the season of their hospitalization, and the vaccination rate was not improved among oxygen-dependent patients.

In this multicenter prospective cohort study, influenza vaccination was associated with a significant reduction

A systematic review of randomized studies found that influenza vaccination reduced COPDE in the same

Clinical outcomes among patients with COPD using home oxygen are described in e-Table 1.

TABLE 3

] Predominant Circulating Influenza Strain and Estimated Adjusted Influenza Vaccine Effectiveness According to Year During the Study Period, 2011 to 2015

Study Year(s)

No. of InfluenzaPositive Hospitalized Patients

Predominant Influenza Strain (No. of Isolates, % of total influenza isolates)

Adjusted OR for Influenza Vaccination and Influenza-Related Hospitalization

Adjusted Influenza Vaccine Effectivenessa (%, 95% CI)

20112015

1,833

Influenza A-H3 (668/1,833; 36.4%)

0.625

37.5% (27.3 to 46.2)

20112012

151

Influenza B-Yamagata (63/151; 41.7%)

0.513

48.7% (19.2 to 67.5)

20122013

583

Influenza A-H3 (372/583; 63.8%)

0.557

44.3% (26.8 to 57.6)

20132014

565

Influenza A-H1 (212/565; 37.5%)

0.517

43.3% (33.8 to 59.6)

20142015

534

Influenza A-H3 (253/534; 47.4%)

0.941

5.9% (–23.6 to 28.3)

a Final adjusted logistic regression models with backward selection included a priori-selected covariates based on univariate analyses. The variables included in the adjusted models for each influenza season are as follows: 2011 to 2015: age categories (16-49, 50-64, 65-75, and > 75 years), long-term care dwelling, heart disease, current smoking, assisted living dwelling, home oxygen use, BMI, previous physician or ED visit for current illness, and > 4 medications at admission. 2011: age, long-term care dwelling, rheumatic disease. 2012 to 2013: age, long-term care dwelling, assisted living dwelling, aboriginal ethnicity, solid organ transplant history, current smoker, home oxygen use, previous physician or ED visit for current illness, > 4 medications at admission. 2013 to 2014: age, long-term care dwelling, current smoker, home oxygen use, previous physician or ED visit for current illness, > 4 medications at admission. 2014 to 2015: age, long-term care dwelling, home oxygen use, assisted living dwelling, previous physician or ED visit for current illness, antiviral medication use prior to admission.

chestjournal.org

73

TABLE 4

] Baseline Characteristics and Outcomes of Hospitalized Patients With COPD and Confirmed Influenza Between 2011 and 2015 According to Influenza Vaccination Status Vaccinated (n ¼ 878)

Unvaccinated n ¼ 612)

16-49 y

14 (1.6%)

27 (4.4%)

50-64 y

131 (14.9%)

178 (29.1%)

65-75 y

232 (26.4%)

151 (24.7%) 256 (41.8%)

Variable

P Value

Age categories < .001

> 75 y

501 (57.1%)

Female sex

458 (52.2%)

313 (51.1%)

.760

BMI (25-40 kg/m2)

402 (45.8%)

278 (45.4%)

.893

Aboriginal ethnicity (n ¼ 114, 6.2% unknown)

8 (0.9%)

15 (2.5%)

Current smoker (n ¼ 86, 4.7% unknown)

235 (26.8%)

266 (43.5%)

< .001

Heart disease

488 (55.6%)

281 (45.9%)

< .001

43 (4.9%)

31 (5.1%)

.518

Kidney or liver disease

147 (16.7%)

102 (16.7%)

.134

Cancer

186 (21.2%)

111 (18.1%)

.350

42 (4.8%)

25 (4.1%)

.686

Living in long-term care facility (n ¼ 15, 0.8% unknown)

105 (12.0%)

32 (5.2%)

< .001

Living in assisted living facility (n ¼ 15, 0.8% unknown)

117 (13.3%)

55 (9.0%)

.028

95 (10.8%)

56 (9.2%)

.573

Physician or ED visit prior to hospitalization for the current illness (n ¼ 32, 1.7% unknown)

289 (32.9%)

201 (32.8%)

.250

No. of medications at admission: > 4

756 (86.1%)

447 (73.0%)

< .001

Antiviral medication use in hospital

635 (72.3%)

402 (65.7%)

.021

H1N1

111 (12.6%)

133 (21.7%)

< .001

H3N2

394 (44.9%)

204 (33.3%)

Unknown

144 (16.4%)

147 (24.0%)

13 (1.5%)

11 (1.8%)

Yamagata

155 (17.7%)

76 (12.4%)

Unknown

61 (6.9%)

41 (6.7%)

67 (7.6%)

63 (10.3%)

Diabetes with end-organ complications

Rheumatic disease

Home oxygen use (n ¼ 1, 0.1% unknown)

.041

Influenza A lineages

Influenza B Lineages Victoria

Use of mechanical ventilation a

.001

.004

Noninvasive positive pressure ventilation

85 (10.7%)

74 (13.3%)

30-d mortality

97 (11.0%)

48 (7.8%)

ICU admission

138 (15.7%)

119 (19.4%)

.025

Length of stay, mean  SD, d

11.0  12.4

11.6  15.0

.410

7 (5, 12)

7 (5, 13)

.410

Length of stay, median (IQR), d

.009 < .001

NIPPV ¼ noninvasive positive pressure ventilation. See Table 2 legend for expansion of other abbreviation. a For NIPPV use, data was missing for 80 patients in vaccinated group, and 57 patients in unvaccinated group.

season but did not show a significant effect on hospitalizations.11 However, a retrospective study among elderly outpatients with chronic lung diseases revealed a significant reduction in hospitalizations for influenza and pneumonia in vaccinated individuals.27 Three additional observational studies evaluated vaccine effectiveness for mortality (one prospective and two

74 Original Research

retrospective).28-30 All three studies found a reduction in all-cause mortality among vaccinated vs unvaccinated individuals, but the results were nonsignificant in the prospective study. We found no significant reduction in mortality associated with vaccination in the present study. The evidence to support a mortality benefit with vaccination remains controversial due to unaccounted

[

155#1 CHEST JANUARY 2019

]

TABLE 5 ]

Risk Factors for Mortality Among Hospitalized Adult Patients With COPD and Confirmed Influenza, Using Adjusted Logistic Regression Models With Backward Selection, 2011 to 2015, Among 1,833 Patients With 204 Deaths

Risk Factors for 30-D Mortality

No. (%) of Deaths

Age > 75 y

144 (70.6)

Heart disease

Adjusted ORa (95% CI) 3.7 (0.4-30.3)

144 (70.6)

2.0 (1.3-3.2)

Cancer

54 (26.5)

1.5 (0.9-2.4)

Admission from long-term care residence

45 (22.1)

2.6 (1.5-4.5)

Current smoking

56 (27.5)

1.6 (0.9-2.6)

Home oxygen use

46 (22.5)

2.9 (1.6-5.1)

Previous physician/ED visit for current illness

77 (37.7)

1.7 (1.1-2.6)

22 (10.8)

1.7 (0.9-3.2)

97 (47.5)

0.9 (0.6-1.4)

BMI < 18.5 kg/m

2

Vaccinated against influenza (protected)

a Final adjusted logistic regression model with backward selection included a priori-selected covariates based on univariate analyses (age categories of 1649, 50-64, 65-75, and > 75 years; vaccination status, heart disease; renal or liver disease; cancer; admission from long-term care facility; oxygen use at baseline; prior physician or ED visit for current illness; BMI categories of < 18.5, 25-29.99, 30-40, and > 40 kg/m2; and number of medications).

frailty selection bias and all-cause mortality end points, which have been suggested to overestimate the effect of influenza vaccination on mortality.31 Crude rates of mortality (11.2%) and critical illness (18.1%) among patients with COPD with influenza in the present study were high. Previous studies reported 2% to 4% influenza-associated mortality in the hospitalized population, and 9% in elderly patients with chronic lung disease, but clinical outcomes in the COPD population have not been well described.16,32,33 Our results suggest that patients with COPD are at higher risk of death if hospitalized with influenza, compared with estimates in the general hospitalized patient population.16,33 Despite guideline recommendations, 34% of hospitalized patients with COPD were not vaccinated in the season of hospitalization. The vaccination rate was not statistically different between oxygen-dependent and nondependent patients (50.8% vs 47.6%). Additional research to identify and mitigate health system barriers TABLE 6 ]

to vaccine uptake could translate to reduced exacerbations and hospitalization and to potential cost savings in this population. In addition to vaccination, antiviral use in the hospital is another potential strategy to reduce influenza-associated complications.34,35 We found that 69% of patients with COPD with influenza received an antiviral agent in the present study. Previous studies have also reported low antiviral prescription rates; one large center reported that only 19.5% of patients with influenza received an antiviral agent.36 Potential reasons for less use of antiviral medication could include clinician doubts about antiviral efficacy beyond the 48 h following symptom onset. However, this 48-h window applies to otherwise healthy adults with uncomplicated influenza.34,35 Subsequent studies of neuraminidase inhibitors have reported a mortality reduction among hospitalized patients with influenza A H1N1pdm09.37,38 The Centers for Disease Control and Prevention recommend use of antiviral medications as early as

Risk Factors for ICU Admission Among Hospitalized Adult Patients With COPD and Confirmed Influenza, Using Adjusted Logistic Regression Models With Backward Selection, 2011 to 2015, Among 1,833 Patients, 332 ICU Admissions

Risk Factors for ICU Admission Diabetes with no end-organ complications Diabetes with end-organ complications Current smoking Home oxygen use

No. (%) of Patients Admitted to ICU

Adjusted ORa (95% CI)

102 (30.7)

1.7 (1.2-2.4)

21 (6.3)

3.5 (1.8-6.8)

158 (47.6)

2.0 (1.4-2.7)

53 (16)

1.8 (1.1-2.9)

a Final adjusted logistic regression model with backward selection included a priori-selected covariates based on univariate analyses (age categories, vaccination status, diabetes, smoking status, admission from assisted living facility, oxygen use at baseline, BMI categories, number of medications, and antiviral medication use).

chestjournal.org

75

possible in the course of hospitalization for any patient with confirmed or suspected influenza.39,40 In the present study, almost one half of the influenza-infected patients admitted to the ICU received the antiviral agent on the same day of ICU admission. Interestingly, this observed clinical practice does not reflect the Centers for Disease Control and Prevention recommendation of prescribing antiviral agents at the onset of hospitalization. However, further research is required in the COPD patient population to specifically address whether antiviral medication initiated at the onset of hospitalization (regardless of the lag time between symptom onset and therapy initiation) could significantly improve clinical outcomes, including mortality, critical illness, and length of hospital stay in patients infected with influenza.

conduct full multiplex PCR analysis on all NP swab samples and therefore cannot confirm the rate of noninfluenza viruses in the case subjects and control subjects. However, given that all patients in the present cohort have COPD (which gives a baseline susceptibility to viral infections) and the lack of vaccines for other respiratory viruses, the likelihood of noninfluenza infections may not differ according to vaccine status. The COPD diagnosis was confirmed by using medical records rather than by spirometry, which could lead to a conservative estimate of COPD cases, as a Canadian study reports that COPD is underdiagnosed in the population.37 We could not adjust our analysis for COPD severity based upon lung function; however, we adjusted for oxygen dependency, which is a credible physiologic surrogate for advanced disease. It is also possible that a selection bias existed, whereby “sicker” patients were more likely to receive their vaccination and therefore more likely to be influenza negative. This theory is evidenced by increased heart disease and oxygen use in the influenza-negative group. These results show worse outcomes among the influenza case subjects, suggesting a limited effect of this bias. Lastly, we did not account for bacterial co-infections in this study; however, previous studies have suggested that clinical outcome severity may not be worse for coinfected patients.38,42

To our knowledge, this analysis is the largest prospective multicenter study of influenza vaccine effectiveness, risk factors, and outcomes among hospitalized patients with COPD. We used a test-negative design, which avoids a potential influenza infection misclassification bias, and the confounding effect from differences in health careseeking behavior among patients in traditional casecontrol and cohort designs.20,41 We collected data during influenza seasons only, such that case subjects and control subjects had equal potential to be exposed to the influenza virus, and second, the case subjects and control subjects were selected from the same population of hospitalized patients with COPD, all of whom sought health care for acute respiratory illness. This approach satisfies two major assumptions of the test-negative design. Lastly, large prospective observational studies are very valuable in providing reliable evidence to support or challenge vaccine recommendations. Given that influenza vaccination is now a guideline-recommended intervention, it would be ethically impossible to conduct new large randomized trials for vaccine effectiveness in this population.

The present study reported a significant reduction in influenza-related hospitalizations associated with seasonal influenza vaccination and identified a high burden of influenza infection among hospitalized patients with COPD. Important clinical care gaps were also identified, including low influenza vaccination rates and delayed initiation of antiviral therapy in this highrisk population of patients with COPD.

The present study has several limitations. One of the assumptions of the test-negative design is that the rate of noninfluenza viruses in the control group should be independent of the influenza vaccine status, to avoid biased vaccine effectiveness estimates.20 We did not

Future research should focus on knowledge translation interventions to increase and sustain vaccine uptake in this highly vulnerable population, and should explore the value of nonvaccine strategies for influenza prevention in highest risk populations.

76 Original Research

Conclusions

[

155#1 CHEST JANUARY 2019

]

Acknowledgments Author contributions: S. M. takes full responsibility for the content of the manuscript, including the data analysis. S. A. M., S. M., L. Y., T. H., M. K. A., and A. A. were involved in the conception and design of the study. S. A. M., M. K. A., and A. A. were responsible for acquisition of data. T. H. and M. E. conducted/supervised the Canadian Immunization Research Network SOS Network central laboratory. S. M., L. Y., L. L., T. H., M. K. A., and S. A. M. analyzed and interpreted the data. F.H. and G. D. S. provided review of the initial analysis. S. M. drafted the manuscript. All authors revised the manuscript critically for important intellectual content, and all authors reviewed and approved the final draft of the manuscript. Financial/nonfinancial disclosures: The authors have reported to CHEST the following: T. H. reports payments from the GlaxoSmithKline (GSK) group of companies, Pfizer, and AbbVie outside the submitted work. M. K. A. reports grant funding from the GSK group of companies, Pfizer, and Sanofi Pasteur. A. C. reports payments from Sanofi outside the submitted work and was previously employed by GSK. G. D. S. was an external consultant at Business & Decision Life Sciences (on behalf of GSK) at the time of the study, is currently employed by the GSK group of companies, and holds shares in the GSK group of companies. M. E. reports payments from the GSK group of companies, the Public Health Agency of Canada, and the Canadian Institutes for Health Research during the conduct of the study. F. H. is employed by the GSK group of companies. S. A. H. reports payments from the GSK group of companies during the conduct of the study and outside the submitted work. B. I. was employed by the GSK group of companies at the time of the study and is now employed by Bayer Inc. J. M. L. reports payments from the GSK group of companies and the Canadian Institutes for Health Research during the conduct of the study and reports payment from the GSK group of companies outside the submitted work. M. L. has received funding honoraria for speaking/ advisory roles for Sanofi, Seqirus, Pfizer, and AstraZeneca. J. E. M. reports payments to her institution from the GSK group of companies and Sanofi Pasteur outside of the submitted work. A. McGeer reports payments to her institution from the GSK group of companies for the conduct of this study and payments from Hoffmann–La Roche and Sanofi Pasteur outside the submitted work. J. P. reports payments from the GSK group of companies, Merck, Roche, and Synthetic Biologics outside the submitted work. M. S. reports payments from the GSK group of companies and Pfizer during the conduct of the study. V. S. was employed by GSK Vaccines at the time of the study and is now employed by Novavax Vaccines and holds shares in the GSK group of companies. S. T. reports payments from the Canadian Institutes for Health Research during the conduct of the study. L. V. reports payments

chestjournal.org

from the GSK group of companies during the conduct of the study. S. A. M. reports payments from the GSK group of companies during the conduct of the study and reports payments from Pfizer, Merck, Novartis and Sanofi outside the submitted work. None declared (S. M., L. L., L. Y., A. A., G. B., W. B., K. G., J. J., K. K., J. L., D. M.-C., A. McCarthy., D. R., D. S., and D. W.). Role of sponsors: The sponsor had no role in the design of the study, the collection and analysis of the data. GSK provided critical review of the manuscript drafts and final version. *Serious Outcomes Surveillance (SOS) Network of the Canadian Immunization Research Network (CIRN) Members: Philippe Lagace-Wiens, MD, St. Boniface Hospital, Winnipeg, MB, Canada; Bruce Light, MD, St. Boniface Hospital, Winnipeg, MB, Canada; Andre Poirier, MD, Centre Intégré Universitaire de santé et services sociaux, Quebec, QC, Canada; Stephanie Smith, MD, University of Alberta Hospital, Edmonton, AB, Canada; and Gregory Taylor, MD, University of Alberta Hospital, Edmonton, AB, Canada. Additional information: The e-Table can be found in the Supplemental Materials section of the online article.

References 1. The Human and Economic Burden of COPD. 2010. https://lungontario.ca/wpcontent/uploads/2017/09/COPDReport_ E_0.pdf Accessed November 30, 2018. 2. Mohan A, Chandra S, Agarwal D, et al. Prevalence of viral infection detected by PCR and RT-PCR in patients with acute exacerbation of COPD: a systematic review. Respirology. 2010;15(3):536-542. 3. Zwaans WA, Mallia P, van Winden ME, Rohde GG. The relevance of respiratory viral infections in the exacerbations of chronic obstructive pulmonary disease—a systematic review. J Clin Virol. 2014;61(2): 181-188. 4. Clark TW, Medina M, Batham S, Curran MD, Parmar S, Nicholson KG. Adults hospitalised with acute respiratory illness rarely have detectable bacteria in the absence of COPD or pneumonia; viral infection predominates in a large prospective UK sample. J Infect. 2014;69(5):507-515.

exacerbations of chronic obstructive pulmonary disease in Kashmir, India, 2010-2012. Influenza Other Respi Viruses. 2015;9(1):40-42. 9. O’Donnell DE, Aaron S, Bourbeau J, et al. Canadian Thoracic Society recommendations for management of chronic obstructive pulmonary disease— 2007 update. Can Respir J. 2007;14(suppl B):5B-32B. 10. Global Initiative for COPD: Guidelines. http://goldcopd.org/. Accessed November 30, 2018. 11. Poole PJ, Chacko E, Wood-Baker RW, Cates CJ. Influenza vaccine for patients with chronic obstructive pulmonary disease. Cochrane Database Syst Rev. 2006;1:CD002733. 12. Wongsurakiat P, Maranetra KN, Wasi C, Kositanont U, Dejsomritrutai W, Charoenratanakul S. Acute respiratory illness in patients with COPD and the effectiveness of influenza vaccination: a randomized controlled study. Chest. 2004;125(6):2011-2020. 13. Howells CH, Tyler LE. Prophylactic use of influenza vaccine in patients with chronic bronchitis. A pilot trial. Lancet. 1961;2(7218):1428-1432. 14. Cimen P, Unlu M, Kirakli C, et al. Should patients with COPD be vaccinated? Respir Care. 2015;60(2):239-243. 15. Garrastazu R, García-Rivero JL, Ruiz M, et al. Prevalence of influenza vaccination in chronic obstructive pulmonary disease patients and impact on the risk of severe exacerbations. Arch Bronconeumol. 2016;52(2):88-95. 16. Mallia P, Johnston SL. Influenza infection and COPD. Int J Chron Obstruct Pulmon Dis. 2007;2(1):55-64. 17. McNeil SA, Andrew MK, Ye L, et al. Interim estimates of 2014/15 influenza vaccine effectiveness in preventing laboratory-confirmed influenza-related hospitalisation from the Serious Outcomes Surveillance Network of the Canadian Immunization Research Network, January 2015. Euro Surveill. 2015;20(5):21024. 18. McNeil S, Shinde V, Andrew M, et al. Interim estimates of 2013/14 influenza clinical severity and vaccine effectiveness in the prevention of laboratory-confirmed influenza-related hospitalisation, Canada, February 2014. Euro Surveill. 2014;19(9).

5. Beckham JD, Cadena A, Lin J, et al. Respiratory viral infections in patients with chronic, obstructive pulmonary disease. J Infect. 2005;50(4):322-330.

19. FluWatch Canada. https://www.canada. ca/en/public-health/services/diseases/fluinfluenza/influenza-surveillance.html. Accessed November 30, 2018.

6. Ko FW, Ip M, Chan PK, et al. Viral etiology of acute exacerbations of COPD in Hong Kong. Chest. 2007;132(3):900908.

20. Jackson ML, Nelson JC. The test-negative design for estimating influenza vaccine effectiveness. Vaccine. 2013;31(17):21652168.

7. Ko FW, Ip M, Chan PK, et al. A 1-year prospective study of the infectious etiology in patients hospitalized with acute exacerbations of COPD. Chest. 2007;131(1):44-52.

21. Foppa IM, Ferdinands JM, Chaves SS, et al. The case test-negative design for studies of the effectiveness of influenza vaccine in inpatient settings. Int J Epidemiol. 2016;45(6):2052-2059.

8. Koul PA, Khan UH, Asad R, et al. Contribution of influenza to acute

22. Ainslie KE, Shi M, Haber M, Orenstein WA. On the bias of estimates of

77

influenza vaccine effectiveness from testnegative studies. Vaccine. 2017;35(52): 7297-7301. 23. Peduzzi P, Concato J, Kemper E, Holford TR, Feinstein AR. A simulation study of the number of events per variable in logistic regression analysis. J Clin Epidemiol. 1996;49(12):1373-1379. 24. U.S. National Library of Medicine, ClinicalTrials.gov database. https:// clinicaltrials.gov/ct2/show/NCT01517191. Accessed November 30, 2018. 25. Zimmerman RK, Nowalk MP, Chung J, et al. 2014-2015 Influenza vaccine effectiveness in the United States by vaccine type. Clin Infect Dis. 2016;63(12): 1564-1573. 26. Centers for Disease Control and Prevention. Seasonal Influenza Vaccine Effectiveness, 2005-2018. https://www.cdc. gov/flu/professionals/vaccination/ effectiveness-studies.htm. Accessed November 30, 2018.

30. Schembri S, Morant S, Winter JH, MacDonald TM. Influenza but not pneumococcal vaccination protects against all-cause mortality in patients with COPD. Thorax. 2009;64(7):567-572. 31. Simonsen L, Taylor RJ, Viboud C, Miller MA, Jackson LA. Mortality benefits of influenza vaccination in elderly people: an ongoing controversy. Lancet Infect Dis. 2007;7(10):658-666. 32. Griffin MR, Coffey CS, Neuzil KM, Mitchel EF, Wright PF, Edwards KM. Winter viruses: influenza- and respiratory syncytial virus-related morbidity in chronic lung disease. Arch Intern Med. 2002;162(11):1229-1236. 33. Loubet P, Samih-Lenzi N, Galtier F, et al. Factors associated with poor outcomes among adults hospitalized for influenza in France: a three-year prospective multicenter study. J Clin Virol. 2016;79: 68-73.

27. Nichol KL, Baken L, Nelson A. Relation between influenza vaccination and outpatient visits, hospitalization, and mortality in elderly persons with chronic lung disease. Ann Intern Med. 1999;130(5):397-403.

34. Aoki FY, Macleod MD, Paggiaro P, et al. Early administration of oral oseltamivir increases the benefits of influenza treatment. J Antimicrob Chemother. 2003;51(1):123-129.

28. Wang CS, Wang ST, Lai CT, Lin LJ, Chou P. Impact of influenza vaccination on major cause-specific mortality. Vaccine. 2007;25(7):1196-1203.

35. Aoki FY, Allen UD, Stiver HG, Evans GA. The use of antiviral drugs for influenza: guidance for practitioners 2012/2013. Can J Infect Dis Med Microbiol. 2012;23(4): e79-e92.

29. Vila-Córcoles A, Ochoa O, de Diego C, et al. Effects of annual influenza vaccination on winter mortality in elderly

78 Original Research

hospitalized patients infected with influenza. Influenza Other Respi Viruses. 2014;8(5):547-548.

people with chronic pulmonary disease. Int J Clin Pract. 2008;62(1):10-17.

36. Herman M, Smieja M, Carruthers S, Loeb M. Oseltamivir use amongst

37. Evans J, Chen Y, Camp PG, Bowie DM, McRae L. Estimating the prevalence of COPD in Canada: reported diagnosis versus measured airflow obstruction. Health Rep. 2014;25(3):3-11. 38. Damasio GA, Pereira LA, Moreira SD, Duarte dos Santos CN, Dalla-Costa LM, Raboni SM. Does virus-bacteria coinfection increase the clinical severity of acute respiratory infection? J Med Virol. 2015;87(9):1456-1461. 39. Centers for Disease Control and Prevention (CDC). Guidance for Antiviral Use in Influenza. 2018. https://www.cdc. gov/flu/professionals/antivirals/summaryclinicians.htm. Accessed November 30, 2018. 40. Stiver HG, Evans GA, Aoki FY, Allen UD, Laverdière M. Guidance on the use of antiviral drugs for influenza in acute care facilities in Canada, 2014-2015. Can J Infect Dis Med Microbiol. 2015;26(1):e5-e8. 41. Fukushima W, Hirota Y. Basic principles of test-negative design in evaluating influenza vaccine effectiveness. Vaccine. 2017;35(36):4796-4800. 42. De Serres G, Lampron N, La Forge J, Rouleau I, Bourbeau J, Weiss K, et al. Importance of viral and bacterial infections in chronic obstructive pulmonary disease exacerbations. J Clin Virol. 2009;46(2):129-133.

[

155#1 CHEST JANUARY 2019

]