American Journal of Infection Control 41 (2013) 1032-7
Contents lists available at ScienceDirect
American Journal of Infection Control
American Journal of Infection Control
journal homepage: www.ajicjournal.org
Major article
Understanding the burden of influenza infection among adults in Canadian hospitals: A comparison of the 2009-2010 pandemic season with the prepandemic and postpandemic seasons Robyn Mitchell MHSc a, *, Geoffrey Taylor MD b, Allison McGeer MD c, Charles Frenette MD d, Kathryn N. Suh MD e, Alice Wong MD f, Kevin Katz MD g, Krista Wilkinson MSc a, Barbara Amihod RN h, Denise Gravel MSc a, and the Canadian Nosocomial Infection Surveillance Program a
Public Health Agency of Canada, Centre for Communicable Diseases and Infection Control, Ottawa, Ontario, Canada Department of Medicine, Division of Infectious Diseases, University of Alberta Hospital, Edmonton, Alberta, Canada c Department of Microbiology, Mount Sinai Hospital, Toronto, Ontario, Canada d Division of Infectious Diseases, Infection Control Service, McGill University Health Centre, Montreal, Quebec, Canada e Department of Infection Prevention and Control, The Ottawa Hospital, Ottawa, Ontario, Canada f Infection Prevention and Control Program, Royal University Hospital, Saskatoon, Saskatchewan, Canada g Department of Infection Prevention and Control, North York General Hospital, Toronto, Ontario, Canada h Department of Infection Prevention and Control, Jewish General Hospital, Montreal, Quebec, Canada b
Key Words: Surveillance Hospitalized adults Influenza Canada Epidemiology Health careeassociated infection
Background: The degree to which the 2009-2010 influenza pandemic season differed from previous and subsequent influenza seasons in Canadian hospitals has not yet been assessed. Methods: Surveillance for laboratory-confirmed influenza among adults in 51 Canadian Nosocomial Infection Surveillance Program hospitals was conducted between November 1, 2006, and May 31, 2011. Inpatient characteristics, treatment, and outcomes of influenza cases in the pandemic season (20092010) were compared with those in the prepandemic (2006-2007 to 2008-2009) and postpandemic (2010-2011) seasons. Results: The incidence of influenza infection was lower in the postpandemic season (1.59/1,000 admissions) compared with the prepandemic seasons (2.00/1,000 admissions; P < .001) and the pandemic season (1.80/1,000 admissions; P < .001). The proportion of cases classified as health caree associated was much smaller during the pandemic season (6.6%) than in either the prepandemic season (23.2%; P < .001) or the postpandemic season (23.6%; P < .001). Inpatients in the pandemic season were significantly younger compared with those in the prepandemic and postpandemic seasons (P < .001). Inpatients in the pandemic season were less likely to have been vaccinated (P < .001), but more likely to be treated with antiviral agents (P < .001), than inpatients in both the prepandemic and postpandemic seasons. Intensive care unit admission was greater during the pandemic season, but there were no significant differences in 30-day mortality among the seasons. Conclusions: Among adult inpatients, the pH1N1 pandemic season differed from seasonal influenza in terms of age, vaccination status, antiviral use, and intensive care unit admission, but not in terms of 30-day mortality. Crown Copyright Ó 2013 Published by Elsevier Inc. on behalf of the Association for Professionals in Infection Control and Epidemiology, Inc. All rights reserved.
Influenza infections cause substantial morbidity and mortality and are a strain on every country’s health care resources. A feature of influenza epidemiology is its seasonality in temperate climates * Address correspondence to Robyn Mitchell, MHSc, 100 Promenade Eglantine Driveway, PL0601E2, Ottawa, Ontario K1A 0K9, Canada. E-mail address: [email protected] (R. Mitchell). Conflict of interest: None to report.
and propensity for year-to-year variability in severity based on genetic changes in the virus. Given these epidemiologic features, most public health systems of developed countries carry out routine laboratory based influenza surveillance to assess the population impact and monitor changes occurring in the virus so as to plan community health services, such as vaccination strategies. Considering that a great deal of the impact, in terms of both clinical consequences and resources required to manage influenza
0196-6553/$36.00 - Crown Copyright Ó 2013 Published by Elsevier Inc. on behalf of the Association for Professionals in Infection Control and Epidemiology, Inc. All rights reserved. http://dx.doi.org/10.1016/j.ajic.2013.06.008
R. Mitchell et al. / American Journal of Infection Control 41 (2013) 1032-7
infection, is borne by the hospital sector, monitoring influenza at the hospital level is important to our understanding of variations in the severity and burden of disease over time.1-4 In 2009 a novel strain of influenza A, (H1N1)pdm09, emerged and rapidly caused a pandemic. Clinical and epidemiologic features of this pandemic have been extensively described in the hospital5-13 and nonhospital12,14 sectors. The present study compared inpatient characteristics, treatment, and outcomes between adults hospitalized with laboratory-confirmed influenza during the pandemic season (2009-2010) and those hospitalized during the prepandemic season (2006-2007 to 2008-2009) and the postpandemic season (2010-2011) in Canadian hospitals. Such information may further our understanding of the course and impact of the influenza virus and lead to better prevention strategies and inpatient management.
1033
Case finding
METHODS
Cases were identified by concurrent or retrospective chart review by trained infection control practitioners. Detailed inpatient questionnaires eliciting information on inpatient laboratory results, inpatient demographics, risk factors, treatment, and outcomes, were completed manually for each case and submitted to the Public Health Agency of Canada. Inpatients who died within 30 days of an initial positive test were reviewed by a CHEC physician to determine whether the death was directly related to influenza, indirectly related to influenza, or unrelated to influenza. Underlying medical conditions that were consistently evaluated in the 2006-2007, 2007-2008, 2009-2010, 2010-2011, and 2011-2012 seasons included chronic lung disease, chronic heart disease, immune suppression, diabetes mellitus, and kidney disease. Data for underlying medical conditions were not available for the 2008-2009 season.
Surveillance network
Data analysis
The Canadian Nosocomial Infection Surveillance Program (CNISP) is a network of 54 largely urban tertiary acute care sentinel hospitals from 10 provinces. The CNISP is a partnership between the Public Health Agency of Canada and the Canadian Hospital Epidemiology Committee (CHEC), a subcommittee of the Association of Medical Microbiology and Infectious Disease (AMMI) Canada. Surveillance of adult influenza at CNISP participating hospitals is generally considered to be within the mandate of hospital infection prevention and control programs and thus does not constitute human research. In most CNISP participating hospitals, this surveillance activity does not require Institutional Review Board review; however, in a minority of hospitals, local policy requires Institutional Review Board review and approval.
Inpatients were divided into 3 cohorts: prepandemic (tested positive for influenza between November 1 and June 30 in the 2006-2007 and 2007-2008 seasons and between November 1, 2008, and April 30, 2009), pandemic (tested positive for influenza between May 1, 2009, and May 31, 2010), and postpandemic (tested positive for influenza between June 1, 2010, and May 31, 2011). The data were examined for consistency and comparability of inpatient characteristics, treatments, and outcomes in the 3 cohorts. Rates of influenza among hospitalized adults were calculated using inpatient admissions. Descriptive statistics were calculated. Differences were assessed for categorical variables using the c2 test. Tests for normality were performed on length of stay and age data. The Kruskall-Wallis test was used to compare the median length of stay and age among the 3 cohorts. Factorial logistic regression was performed to identify differences in inpatient characteristics, treatments, and outcomes by cohort. The pandemic cohort served as the reference. Odds ratios (OR) are 95% confidence intervals (CI) are reported. P values reflect a 2-tailed a level of .05. Missing data and unassessable responses were removed from all calculations. All statistical analyses were performed using Stata version 11 (StataCorp, College Station, TX).
Surveillance period Between 2006 and 2008, the CNISP conducted surveillance of laboratory-confirmed influenza among hospitalized inpatients age 16 years and older during the traditional influenza season (November-June). After emergence of the pH1N1 influenza virus in April 2009, the surveillance program was expanded to include yearround surveillance, which was continued into the 2010-2011 influenza season.
Case definition During influenza season, clinical practice guidelines recommend testing adults of any age who are hospitalized with fever and respiratory symptoms, including community-acquired pneumonia.15 An influenza case was defined as any adult (age 16 years) with a positive influenza laboratory test result (confirmed by rapid antigen test, reverse-transcriptase polymerase chain reaction [RT-PCR], or viral culture) from a specimen collected during the surveillance period who was admitted to a CNISP participating hospital. Patients seen in outpatient clinics, ambulatory care, and emergency departments whose visits did not result in hospitalization were excluded. A health careeassociated case was defined as a case meeting 1 of the following criteria: (1) onset of symptoms more than 96 hours after admission or (2) onset of symptoms 24-96 hours after admission, with illness acquired in the hospital in the judgment of the infection control professional, or (3) readmission with a positive test within 96 hours after discharge.
RESULTS A total of 2,868 inpatients with laboratory-confirmed influenza from 51 CNISP hospitals were reported between November 1, 2006, and May 31, 2011. A total of 629 cases were reported in the prepandemic season (from 18 reporting hospitals), 1,132 cases were reported in the pandemic season (from 43 reporting hospitals), and 1,107 cases were reported in the postpandemic season (from 35 reporting hospitals) (Table 1). The rate of laboratory-confirmed influenza-associated hospitalization was significantly lower in the postpandemic season (1.59/1,000 inpatient admissions) compared with the pandemic season (1.80/1,000 inpatient admissions; P ¼ .004) and the prepandemic season (2.00/1,000 inpatient admissions; P < .001), although there was no significant difference in the influenza infection rate in the prepandemic and pandemic seasons (P ¼ .35). The proportion of health careeassociated cases was significantly lower in the pandemic season (6.6%) compared with the prepandemic season (23.2%; P < .001) and the postpandemic season (23.6%; P < .001). The proportion of health careeassociated cases did not differ significantly between the prepandemic and postpandemic seasons (P ¼ .84). The proportion of health caree associated cases acquired in an acute-care facility instead of
1034
R. Mitchell et al. / American Journal of Infection Control 41 (2013) 1032-7
Table 1 Patient characteristics, treatments, and outcomes of laboratory-confirmed influenza among adults hospitalized in CNISP hospitals, 2006-2011
Participating hospitals, n Influenza cases, n Influenza A, n (%) Influenza B, n (%) Unknown/not typed, n (%) Age group, n (%) 16-24 years 25-49 years 50-64 years 65þ years Age, years Mean (SD) Median Male sex, n (%) Underlying medical condition, n (%) Any condition Chronic heart disease Diabetes Immune suppressed Chronic lung disease Chronic kidney disease Received vaccine Yes, n (%) Received antibiotics Yes, n (%) Received antivirals Yes, n (%) If yes, antiviral received, n (%) Oseltamivir Zanamivir Amantadine ICU admission Yes, n (%) Intubation Yes, n (%) 30-day outcome Discharged, n (%) Remains hospitalized, n (%) Died, n (%) Death attributable to influenza, n (%)* Relationship of influenza to death, n (%) Primary cause Contributing cause Not related Unable to determine Source of infection Community-associated, n (%) Health careeassociated, n (%) Source of health careeassociated infection Acute-care facility, n (%) Long-term care facility, n (%) For community-associated cases only Length of stay, days Mean (SD) Median
Prepandemic influenza cohort (2006-2009)
Pandemic influenza cohort (2009-2010)
Postpandemic influenza cohort (2010-2011)
18 629 444 (70.6) 182 (28.9) 3 (0.5)
43 1,132 1,128 (99.6) 4 (0.4) 0 (0)
35 1,107 1,015 (91.7) 89 (8.0) 3 (0.3)
29 120 100 380
(4.6) (19.1) (15.9) (60.4)
133 (11.7) 476 (42.0) 371 (32.8) 152 13.4)
23 181 158 745
(2.1) (16.4) (14.3) (67.3)
66.1 (21.0) 72 315 (50.1)
46.5 (16.4) 48 526 (46.5)
69.7 (19.9) 75 509 (46.0)
446 (70.9) 115 (18.3) 89 (14.1) 74 (11.8) 137 (21.8) 45 (7.2) N ¼ 249 130 (52.2) N ¼ 600 394 (65.7) N ¼ 610 243 (39.8)
902 (79.7) 158 (14.0) 168 (14.8) 189 (16.7) 409 (36.1) 74 (6.5) N ¼ 485 103 (21.2) N ¼ 1,112 855 (76.9) N ¼ 1,113 997 (89.6)
956 (86.4) 392 (35.4) 241 (21.8) 184 (16.6) 337 (30.4) 124 (11.2) N ¼ 391 179 (45.8) N ¼ 1,054 791 (75.0) N ¼ 1,092 873 (79.9)
232 (95.5) 5 (2.1) 6 (2.5) N ¼ 603 76 (12.6) N ¼ 613 57 (9.3) N ¼ 628 478 (76.1) 97 (15.4) 53 (8.4) 25 (4.0)
984 (98.7) 10 (1.0) 0 (0) N ¼ 1,125 289 (25.7) N ¼ 1,127 245 (21.7) N ¼ 1,132 978 (86.4) 82 (7.2) 72 (6.4) 62 (5.5)
870 (99.7) 2 (0.2) 0 (0) N ¼ 1,093 138 (12.6) N ¼ 1,110 108 (9.7) N ¼ 1,107 904 (81.7) 129 (11.7) 74 (6.7) 50 (4.5)
10 (18.9) 15 (28.3) 22 (41.5) 6 (11.3) N ¼ 617 474 (76.8) 143 (23.2) N ¼ 143 67 (46.9) 76 (53.1)
27 (37.5) 35 (48.6) 4 (5.6) 6 (8.3) N ¼ 1,000 934 (93.4) 66 (6.6) N ¼ 66 41 (62.1) 25 (37.9)
14 (18.9) 36 (48.6) 16 (21.6) 8 (10.8) N ¼ 1,089 832 (76.4) 257 (23.6) N ¼ 257 82 (31.9) 175 (68.1)
9.5 (15.4) 5
7.6 (8.3) 5
7.3 (8.1) 5
*Attributable influenza death: influenza was the primary or contributing cause of death.
a long-term care facility was significantly higher during the pandemic season (62.1%) compared with the prepandemic season (46.9%; P ¼ .04) and the postpandemic season (31.9%; P < .001). As expected, significantly fewer inpatients were infected with influenza B during the pandemic season (0.04%) compared with the prepandemic season (28.9%; P < .001) and the postpandemic season (8.0%; P < .001). For community-associated cases, the median length of stay did not differ between seasons (P ¼ .29) (Table 2). Inpatients in the pandemic season were significantly younger than those in the prepandemic and postpandemic seasons (P < .001) (Table 1). No significant differences in sex distribution were observed among the 3 seasons.
Some differences in underlying medical conditions were observed among the seasons. When adjusted for age, inpatients in the postpandemic season were significantly more likely to have chronic heart and kidney disease compared with inpatients in the pandemic season, whereas inpatients in the pandemic season were significantly more likely to have pulmonary disease compared with those in the prepandemic and postpandemic seasons (Table 2). Prevention and treatment of influenza also differed among the seasons. Inpatients in the pandemic season were less likely to have been vaccinated but more likely to be treated with antibiotics and antivirals compared with inpatients in both the prepandemic and postpandemic seasons (Table 2). More severe outcomes were reported among inpatients in the pandemic season. These inpatients
R. Mitchell et al. / American Journal of Infection Control 41 (2013) 1032-7 Table 2 Factorial logistic regression of the pandemic season compared with the prepandemic and postpandemic seasons
Any underlying medical condition Prepandemic Pandemic Postpandemic Chronic heart disease Prepandemic Pandemic Postpandemic Chronic lung disease Prepandemic Pandemic Postpandemic Diabetes Prepandemic Pandemic Postpandemic Immune suppressed Prepandemic Pandemic Postpandemic Chronic kidney disease Prepandemic Pandemic Postpandemic Received vaccine Prepandemic Pandemic Postpandemic Received antibiotics Prepandemic Pandemic Postpandemic Received antivirals Prepandemic Pandemic Postpandemic ICU admission Prepandemic Pandemic Postpandemic Intubation/mechanical ventilation Prepandemic Pandemic Postpandemic Death at 30 days Prepandemic Pandemic Postpandemic Death attributable to influenza Prepandemic Pandemic Postpandemic
Age-adjusted OR
95% CI
P value
0.79 1 0.85
0.58-1.06 Reference 0.66-1.10
.117
0.72 1 1.79
0.54-0.98 Reference 1.40-2.29
.034
0.51 1 0.77
0.40-0.65 Reference 0.63-0.95
Contents lists available at ScienceDirect
American Journal of Infection Control
American Journal of Infection Control
journal homepage: www.ajicjournal.org
Major article
Understanding the burden of influenza infection among adults in Canadian hospitals: A comparison of the 2009-2010 pandemic season with the prepandemic and postpandemic seasons Robyn Mitchell MHSc a, *, Geoffrey Taylor MD b, Allison McGeer MD c, Charles Frenette MD d, Kathryn N. Suh MD e, Alice Wong MD f, Kevin Katz MD g, Krista Wilkinson MSc a, Barbara Amihod RN h, Denise Gravel MSc a, and the Canadian Nosocomial Infection Surveillance Program a
Public Health Agency of Canada, Centre for Communicable Diseases and Infection Control, Ottawa, Ontario, Canada Department of Medicine, Division of Infectious Diseases, University of Alberta Hospital, Edmonton, Alberta, Canada c Department of Microbiology, Mount Sinai Hospital, Toronto, Ontario, Canada d Division of Infectious Diseases, Infection Control Service, McGill University Health Centre, Montreal, Quebec, Canada e Department of Infection Prevention and Control, The Ottawa Hospital, Ottawa, Ontario, Canada f Infection Prevention and Control Program, Royal University Hospital, Saskatoon, Saskatchewan, Canada g Department of Infection Prevention and Control, North York General Hospital, Toronto, Ontario, Canada h Department of Infection Prevention and Control, Jewish General Hospital, Montreal, Quebec, Canada b
Key Words: Surveillance Hospitalized adults Influenza Canada Epidemiology Health careeassociated infection
Background: The degree to which the 2009-2010 influenza pandemic season differed from previous and subsequent influenza seasons in Canadian hospitals has not yet been assessed. Methods: Surveillance for laboratory-confirmed influenza among adults in 51 Canadian Nosocomial Infection Surveillance Program hospitals was conducted between November 1, 2006, and May 31, 2011. Inpatient characteristics, treatment, and outcomes of influenza cases in the pandemic season (20092010) were compared with those in the prepandemic (2006-2007 to 2008-2009) and postpandemic (2010-2011) seasons. Results: The incidence of influenza infection was lower in the postpandemic season (1.59/1,000 admissions) compared with the prepandemic seasons (2.00/1,000 admissions; P < .001) and the pandemic season (1.80/1,000 admissions; P < .001). The proportion of cases classified as health caree associated was much smaller during the pandemic season (6.6%) than in either the prepandemic season (23.2%; P < .001) or the postpandemic season (23.6%; P < .001). Inpatients in the pandemic season were significantly younger compared with those in the prepandemic and postpandemic seasons (P < .001). Inpatients in the pandemic season were less likely to have been vaccinated (P < .001), but more likely to be treated with antiviral agents (P < .001), than inpatients in both the prepandemic and postpandemic seasons. Intensive care unit admission was greater during the pandemic season, but there were no significant differences in 30-day mortality among the seasons. Conclusions: Among adult inpatients, the pH1N1 pandemic season differed from seasonal influenza in terms of age, vaccination status, antiviral use, and intensive care unit admission, but not in terms of 30-day mortality. Crown Copyright Ó 2013 Published by Elsevier Inc. on behalf of the Association for Professionals in Infection Control and Epidemiology, Inc. All rights reserved.
Influenza infections cause substantial morbidity and mortality and are a strain on every country’s health care resources. A feature of influenza epidemiology is its seasonality in temperate climates * Address correspondence to Robyn Mitchell, MHSc, 100 Promenade Eglantine Driveway, PL0601E2, Ottawa, Ontario K1A 0K9, Canada. E-mail address: [email protected] (R. Mitchell). Conflict of interest: None to report.
and propensity for year-to-year variability in severity based on genetic changes in the virus. Given these epidemiologic features, most public health systems of developed countries carry out routine laboratory based influenza surveillance to assess the population impact and monitor changes occurring in the virus so as to plan community health services, such as vaccination strategies. Considering that a great deal of the impact, in terms of both clinical consequences and resources required to manage influenza
0196-6553/$36.00 - Crown Copyright Ó 2013 Published by Elsevier Inc. on behalf of the Association for Professionals in Infection Control and Epidemiology, Inc. All rights reserved. http://dx.doi.org/10.1016/j.ajic.2013.06.008
R. Mitchell et al. / American Journal of Infection Control 41 (2013) 1032-7
infection, is borne by the hospital sector, monitoring influenza at the hospital level is important to our understanding of variations in the severity and burden of disease over time.1-4 In 2009 a novel strain of influenza A, (H1N1)pdm09, emerged and rapidly caused a pandemic. Clinical and epidemiologic features of this pandemic have been extensively described in the hospital5-13 and nonhospital12,14 sectors. The present study compared inpatient characteristics, treatment, and outcomes between adults hospitalized with laboratory-confirmed influenza during the pandemic season (2009-2010) and those hospitalized during the prepandemic season (2006-2007 to 2008-2009) and the postpandemic season (2010-2011) in Canadian hospitals. Such information may further our understanding of the course and impact of the influenza virus and lead to better prevention strategies and inpatient management.
1033
Case finding
METHODS
Cases were identified by concurrent or retrospective chart review by trained infection control practitioners. Detailed inpatient questionnaires eliciting information on inpatient laboratory results, inpatient demographics, risk factors, treatment, and outcomes, were completed manually for each case and submitted to the Public Health Agency of Canada. Inpatients who died within 30 days of an initial positive test were reviewed by a CHEC physician to determine whether the death was directly related to influenza, indirectly related to influenza, or unrelated to influenza. Underlying medical conditions that were consistently evaluated in the 2006-2007, 2007-2008, 2009-2010, 2010-2011, and 2011-2012 seasons included chronic lung disease, chronic heart disease, immune suppression, diabetes mellitus, and kidney disease. Data for underlying medical conditions were not available for the 2008-2009 season.
Surveillance network
Data analysis
The Canadian Nosocomial Infection Surveillance Program (CNISP) is a network of 54 largely urban tertiary acute care sentinel hospitals from 10 provinces. The CNISP is a partnership between the Public Health Agency of Canada and the Canadian Hospital Epidemiology Committee (CHEC), a subcommittee of the Association of Medical Microbiology and Infectious Disease (AMMI) Canada. Surveillance of adult influenza at CNISP participating hospitals is generally considered to be within the mandate of hospital infection prevention and control programs and thus does not constitute human research. In most CNISP participating hospitals, this surveillance activity does not require Institutional Review Board review; however, in a minority of hospitals, local policy requires Institutional Review Board review and approval.
Inpatients were divided into 3 cohorts: prepandemic (tested positive for influenza between November 1 and June 30 in the 2006-2007 and 2007-2008 seasons and between November 1, 2008, and April 30, 2009), pandemic (tested positive for influenza between May 1, 2009, and May 31, 2010), and postpandemic (tested positive for influenza between June 1, 2010, and May 31, 2011). The data were examined for consistency and comparability of inpatient characteristics, treatments, and outcomes in the 3 cohorts. Rates of influenza among hospitalized adults were calculated using inpatient admissions. Descriptive statistics were calculated. Differences were assessed for categorical variables using the c2 test. Tests for normality were performed on length of stay and age data. The Kruskall-Wallis test was used to compare the median length of stay and age among the 3 cohorts. Factorial logistic regression was performed to identify differences in inpatient characteristics, treatments, and outcomes by cohort. The pandemic cohort served as the reference. Odds ratios (OR) are 95% confidence intervals (CI) are reported. P values reflect a 2-tailed a level of .05. Missing data and unassessable responses were removed from all calculations. All statistical analyses were performed using Stata version 11 (StataCorp, College Station, TX).
Surveillance period Between 2006 and 2008, the CNISP conducted surveillance of laboratory-confirmed influenza among hospitalized inpatients age 16 years and older during the traditional influenza season (November-June). After emergence of the pH1N1 influenza virus in April 2009, the surveillance program was expanded to include yearround surveillance, which was continued into the 2010-2011 influenza season.
Case definition During influenza season, clinical practice guidelines recommend testing adults of any age who are hospitalized with fever and respiratory symptoms, including community-acquired pneumonia.15 An influenza case was defined as any adult (age 16 years) with a positive influenza laboratory test result (confirmed by rapid antigen test, reverse-transcriptase polymerase chain reaction [RT-PCR], or viral culture) from a specimen collected during the surveillance period who was admitted to a CNISP participating hospital. Patients seen in outpatient clinics, ambulatory care, and emergency departments whose visits did not result in hospitalization were excluded. A health careeassociated case was defined as a case meeting 1 of the following criteria: (1) onset of symptoms more than 96 hours after admission or (2) onset of symptoms 24-96 hours after admission, with illness acquired in the hospital in the judgment of the infection control professional, or (3) readmission with a positive test within 96 hours after discharge.
RESULTS A total of 2,868 inpatients with laboratory-confirmed influenza from 51 CNISP hospitals were reported between November 1, 2006, and May 31, 2011. A total of 629 cases were reported in the prepandemic season (from 18 reporting hospitals), 1,132 cases were reported in the pandemic season (from 43 reporting hospitals), and 1,107 cases were reported in the postpandemic season (from 35 reporting hospitals) (Table 1). The rate of laboratory-confirmed influenza-associated hospitalization was significantly lower in the postpandemic season (1.59/1,000 inpatient admissions) compared with the pandemic season (1.80/1,000 inpatient admissions; P ¼ .004) and the prepandemic season (2.00/1,000 inpatient admissions; P < .001), although there was no significant difference in the influenza infection rate in the prepandemic and pandemic seasons (P ¼ .35). The proportion of health careeassociated cases was significantly lower in the pandemic season (6.6%) compared with the prepandemic season (23.2%; P < .001) and the postpandemic season (23.6%; P < .001). The proportion of health careeassociated cases did not differ significantly between the prepandemic and postpandemic seasons (P ¼ .84). The proportion of health caree associated cases acquired in an acute-care facility instead of
1034
R. Mitchell et al. / American Journal of Infection Control 41 (2013) 1032-7
Table 1 Patient characteristics, treatments, and outcomes of laboratory-confirmed influenza among adults hospitalized in CNISP hospitals, 2006-2011
Participating hospitals, n Influenza cases, n Influenza A, n (%) Influenza B, n (%) Unknown/not typed, n (%) Age group, n (%) 16-24 years 25-49 years 50-64 years 65þ years Age, years Mean (SD) Median Male sex, n (%) Underlying medical condition, n (%) Any condition Chronic heart disease Diabetes Immune suppressed Chronic lung disease Chronic kidney disease Received vaccine Yes, n (%) Received antibiotics Yes, n (%) Received antivirals Yes, n (%) If yes, antiviral received, n (%) Oseltamivir Zanamivir Amantadine ICU admission Yes, n (%) Intubation Yes, n (%) 30-day outcome Discharged, n (%) Remains hospitalized, n (%) Died, n (%) Death attributable to influenza, n (%)* Relationship of influenza to death, n (%) Primary cause Contributing cause Not related Unable to determine Source of infection Community-associated, n (%) Health careeassociated, n (%) Source of health careeassociated infection Acute-care facility, n (%) Long-term care facility, n (%) For community-associated cases only Length of stay, days Mean (SD) Median
Prepandemic influenza cohort (2006-2009)
Pandemic influenza cohort (2009-2010)
Postpandemic influenza cohort (2010-2011)
18 629 444 (70.6) 182 (28.9) 3 (0.5)
43 1,132 1,128 (99.6) 4 (0.4) 0 (0)
35 1,107 1,015 (91.7) 89 (8.0) 3 (0.3)
29 120 100 380
(4.6) (19.1) (15.9) (60.4)
133 (11.7) 476 (42.0) 371 (32.8) 152 13.4)
23 181 158 745
(2.1) (16.4) (14.3) (67.3)
66.1 (21.0) 72 315 (50.1)
46.5 (16.4) 48 526 (46.5)
69.7 (19.9) 75 509 (46.0)
446 (70.9) 115 (18.3) 89 (14.1) 74 (11.8) 137 (21.8) 45 (7.2) N ¼ 249 130 (52.2) N ¼ 600 394 (65.7) N ¼ 610 243 (39.8)
902 (79.7) 158 (14.0) 168 (14.8) 189 (16.7) 409 (36.1) 74 (6.5) N ¼ 485 103 (21.2) N ¼ 1,112 855 (76.9) N ¼ 1,113 997 (89.6)
956 (86.4) 392 (35.4) 241 (21.8) 184 (16.6) 337 (30.4) 124 (11.2) N ¼ 391 179 (45.8) N ¼ 1,054 791 (75.0) N ¼ 1,092 873 (79.9)
232 (95.5) 5 (2.1) 6 (2.5) N ¼ 603 76 (12.6) N ¼ 613 57 (9.3) N ¼ 628 478 (76.1) 97 (15.4) 53 (8.4) 25 (4.0)
984 (98.7) 10 (1.0) 0 (0) N ¼ 1,125 289 (25.7) N ¼ 1,127 245 (21.7) N ¼ 1,132 978 (86.4) 82 (7.2) 72 (6.4) 62 (5.5)
870 (99.7) 2 (0.2) 0 (0) N ¼ 1,093 138 (12.6) N ¼ 1,110 108 (9.7) N ¼ 1,107 904 (81.7) 129 (11.7) 74 (6.7) 50 (4.5)
10 (18.9) 15 (28.3) 22 (41.5) 6 (11.3) N ¼ 617 474 (76.8) 143 (23.2) N ¼ 143 67 (46.9) 76 (53.1)
27 (37.5) 35 (48.6) 4 (5.6) 6 (8.3) N ¼ 1,000 934 (93.4) 66 (6.6) N ¼ 66 41 (62.1) 25 (37.9)
14 (18.9) 36 (48.6) 16 (21.6) 8 (10.8) N ¼ 1,089 832 (76.4) 257 (23.6) N ¼ 257 82 (31.9) 175 (68.1)
9.5 (15.4) 5
7.6 (8.3) 5
7.3 (8.1) 5
*Attributable influenza death: influenza was the primary or contributing cause of death.
a long-term care facility was significantly higher during the pandemic season (62.1%) compared with the prepandemic season (46.9%; P ¼ .04) and the postpandemic season (31.9%; P < .001). As expected, significantly fewer inpatients were infected with influenza B during the pandemic season (0.04%) compared with the prepandemic season (28.9%; P < .001) and the postpandemic season (8.0%; P < .001). For community-associated cases, the median length of stay did not differ between seasons (P ¼ .29) (Table 2). Inpatients in the pandemic season were significantly younger than those in the prepandemic and postpandemic seasons (P < .001) (Table 1). No significant differences in sex distribution were observed among the 3 seasons.
Some differences in underlying medical conditions were observed among the seasons. When adjusted for age, inpatients in the postpandemic season were significantly more likely to have chronic heart and kidney disease compared with inpatients in the pandemic season, whereas inpatients in the pandemic season were significantly more likely to have pulmonary disease compared with those in the prepandemic and postpandemic seasons (Table 2). Prevention and treatment of influenza also differed among the seasons. Inpatients in the pandemic season were less likely to have been vaccinated but more likely to be treated with antibiotics and antivirals compared with inpatients in both the prepandemic and postpandemic seasons (Table 2). More severe outcomes were reported among inpatients in the pandemic season. These inpatients
R. Mitchell et al. / American Journal of Infection Control 41 (2013) 1032-7 Table 2 Factorial logistic regression of the pandemic season compared with the prepandemic and postpandemic seasons
Any underlying medical condition Prepandemic Pandemic Postpandemic Chronic heart disease Prepandemic Pandemic Postpandemic Chronic lung disease Prepandemic Pandemic Postpandemic Diabetes Prepandemic Pandemic Postpandemic Immune suppressed Prepandemic Pandemic Postpandemic Chronic kidney disease Prepandemic Pandemic Postpandemic Received vaccine Prepandemic Pandemic Postpandemic Received antibiotics Prepandemic Pandemic Postpandemic Received antivirals Prepandemic Pandemic Postpandemic ICU admission Prepandemic Pandemic Postpandemic Intubation/mechanical ventilation Prepandemic Pandemic Postpandemic Death at 30 days Prepandemic Pandemic Postpandemic Death attributable to influenza Prepandemic Pandemic Postpandemic
Age-adjusted OR
95% CI
P value
0.79 1 0.85
0.58-1.06 Reference 0.66-1.10
.117
0.72 1 1.79
0.54-0.98 Reference 1.40-2.29
.034
0.51 1 0.77
0.40-0.65 Reference 0.63-0.95
