Accepted Manuscript Respiratory syncytial virus in adults with severe acute respiratory illness in a high HIV prevalence setting Jocelyn Moyes, Sibongile Walaza, Marthi Pretorius, Michelle Groome, Anne von Gottberg, Nicole Wolter, Sumayya Haffejee, Ebrahim Variava, Adam L. Cohen, Stefano Tempia, Kathleen Kahn, Halima Dawood, Marietjie Venter, Cheryl Cohen, Shabir A. Madhi PII:
S0163-4453(17)30225-6
DOI:
10.1016/j.jinf.2017.06.007
Reference:
YJINF 3947
To appear in:
Journal of Infection
Received Date: 23 April 2017 Revised Date:
22 June 2017
Accepted Date: 24 June 2017
Please cite this article as: Moyes J, Walaza S, Pretorius M, Groome M, von Gottberg A, Wolter N, Haffejee S, Variava E, Cohen AL, Tempia S, Kahn K, Dawood H, Venter M, Cohen C, Madhi SA, South African Severe Acute Respiratory Illness (SARI) Surveillance Group, Respiratory syncytial virus in adults with severe acute respiratory illness in a high HIV prevalence setting, Journal of Infection (2017), doi: 10.1016/j.jinf.2017.06.007. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Respiratory syncytial virus in adults with severe acute respiratory illness in a high HIV prevalence setting. Authors: Jocelyn Moyes1,2, Sibongile Walaza1,
2
, Marthi Pretorius3 , Michelle Groome4, Anne von
RI PT
Gottberg1, Nicole Wolter1, Sumayya Haffejee5, Ebrahim Variava6, Adam L Cohen7, Stefano Tempia1,8 , Kathleen Kahn9 , Halima Dawood10 , Marietjie Venter3, Cheryl Cohen1,2 Shabir A Madhi1,4for the South African Severe Acute Respiratory Illness (SARI) Surveillance Group 1
Centre for Respiratory Diseases and Meningitis, National Institute for
SC
Affiliations:
M AN U
Communicable Diseases of the National Health Laboratory Service, Johannesburg, South Africa; 2School of Public Health, Health Sciences Faculty, University of the Witwatersrand, Johannesburg, South Africa. 3Department of Medical Virology, University of Pretoria, 4
Medical Research Council: Respiratory and Meningeal Pathogens Research Unit and
Department of Science and Technology/National Research Foundation: Vaccine Preventable
TE D
Diseases, University of the Witwatersrand; Johannesburg; South Africa;
5
National health
Laboratory Service; 6Klerksdorp-Tshepong Hospital Complex and The University of the
EP
Witwatersrand; 7Strategic Information Group, Expanded Programme on Immunization, Department of Immunization, Vaccines and Biologicals, World Health Organization
8
AC C
Influenza Program, Centers for Disease Control and Prevention, Pretoria, South Africa and Influenza Division, Centers for Disease Control and Prevention, Atlanta, Georgia, United States of America;
9
MRC/Wits Rural Public Health and Health Transitions Research Unit
(Agincourt), School of Public Health, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa Centre for Global Health Research and Umeå University, Umeå, Sweden and INDEPTH network Accra Ghana;
10
Pietermaritzburg
Hospital Complex and University of KwaZulu-Natal, Pietermaritzburg, South Africa;+
ACCEPTED MANUSCRIPT Corresponding author: Jocelyn Moyes, Centre for Respiratory Diseases and Meningitis, National Institute for Communicable Diseases, Private Bag X4, Sandringham, 2131, Gauteng, South Africa, Telephone: 27 11 386 6410. Fax: 27 11 882 9979. E-mail:
AC C
EP
TE D
M AN U
SC
Word count: Text 3206 words, Abstract 197
RI PT
[email protected]
ACCEPTED MANUSCRIPT 1
Respiratory syncytial virus in adults with severe acute respiratory illness in a high HIV
2
prevalence setting.
3 Authors:
5
Jocelyn Moyes1,2, Sibongile Walaza1, 2 , Marthi Pretorius3 , Michelle Groome4, Anne von Gottberg1,
6
Nicole Wolter1, Sumayya Haffejee5, Ebrahim Variava6, Adam L Cohen7, Stefano Tempia1,8 , Kathleen
7
Kahn9 , Halima Dawood10 , Marietjie Venter3, Cheryl Cohen1,2 Shabir A Madhi1,4for the South African
8
Severe Acute Respiratory Illness (SARI) Surveillance Group
9
Affiliations: 1Centre for Respiratory Diseases and Meningitis, National Institute for Communicable
10
Diseases of the National Health Laboratory Service, Johannesburg, South Africa; 2School of Public
11
Health, Health Sciences Faculty, University of the Witwatersrand, Johannesburg, South Africa.
12
3
13
and Meningeal Pathogens Research Unit and Department of Science and Technology/National
14
Research Foundation: Vaccine Preventable Diseases, University of the Witwatersrand; Johannesburg;
15
South Africa; 5 National health Laboratory Service; 6Klerksdorp-Tshepong Hospital Complex and The
16
University of the Witwatersrand;
17
Immunization, Department of Immunization, Vaccines and Biologicals, World Health Organization
18
Influenza Program, Centers for Disease Control and Prevention, Pretoria, South Africa and Influenza
19
Division, Centers for Disease Control and Prevention, Atlanta, Georgia, United States of America;
20
MRC/Wits Rural Public Health and Health Transitions Research Unit (Agincourt), School of Public
21
Health, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa
22
Centre for Global Health Research and Umeå University, Umeå, Sweden and INDEPTH network
23
Accra Ghana;
24
Pietermaritzburg, South Africa;+
25
Corresponding author: Jocelyn Moyes, Centre for Respiratory Diseases and Meningitis, National
26
Institute for Communicable Diseases, Private Bag X4, Sandringham, 2131, Gauteng, South Africa,
27
Telephone: 27 11 386 6410. Fax: 27 11 882 9979. E-mail: [email protected]
M AN U
SC
RI PT
4
TE D
Department of Medical Virology, University of Pretoria, 4Medical Research Council: Respiratory
Strategic Information Group, Expanded Programme on 8
9
AC C
EP
7
10
Pietermaritzburg Hospital Complex and University of KwaZulu-Natal,
1
ACCEPTED MANUSCRIPT 28
29
Word count: Text 3479 words, Abstract 197
30
AC C
EP
TE D
M AN U
SC
RI PT
31
2
ACCEPTED MANUSCRIPT Abstract
33
Background
34
There are limited data on the epidemiology of respiratory syncytial virus (RSV) illness in HIV-
35
infected adults or the elderly in Africa. We studied the epidemiology of RSV-associated severe acute
36
respiratory illness (SARI) hospitalizations in adults in South Africa from 2009 through 2013.
37
Methods
38
Individuals admitted to sentinel surveillance hospitals were investigated by respiratory tract swabs for
39
RSV, using a multiplex real-time polymerase chain reaction assay. The incidence of RSV-associated
40
SARI was calculated for the one site with population denominators.
41
Results
42
Of 7796 participants investigated, 329 (4%) tested positive for RSV. On multivariable analysis, HIV-
43
infected individuals with RSV-associated SARI had greater odds of being in the age groups 18-44 and
44
45-64 years (odd ratios (OR) 26.3; 95% confidence interval (CI) 6.2-112.1 and OR 11.4; 95% CI 2.6-
45
50.0) compared with those ≥65 years and being female (OR 2.7; 95% CI 1.4-5.4). The relative risk of
46
hospitalization with RSV-associated SARI was 12 to 18 times higher in HIV infected individual
47
compared to that of HIV-uninfected.
48
Conclusion
49
The incidence of RSV-associated SARI was higher in HIV-infected individuals and those aged 65
50
years and older. Further studies are warranted to describe the disease association of RSV detected in
51
adults with SARI.
SC
M AN U
TE D
EP
AC C
52
RI PT
32
3
ACCEPTED MANUSCRIPT 53 Introduction
55
There are limited data on the epidemiology of respiratory syncytial virus (RSV) infection in adults
56
and older children in high HIV prevalence settings like South Africa. In high income countries with a
57
lower HIV burden, the burden of hospitalization with RSV in the elderly has been described as similar
58
to that of influenza (1,2,3,4). Also, RSV infection has been reported to be more severe in the elderly,
59
people living in long term care facilities, immune compromised individuals (specifically those with
60
stem cell and lung transplants) and individuals receiving chemotherapy (4,5,6). Other risk factors in
61
adults include underlying cardiac disease and chronic respiratory illness (7). In a prospective hospital-
62
based surveillance programme in the United States amongst elderly (≥65 years), RSV was identified
63
in 11% of pneumonia episodes, 11% of chronic obstructive pulmonary disease episodes, 5% of
64
congestive heart failure events and 7% of asthma cases (8).
65
In South Africa the highest HIV prevalence occurs among adults aged 15-49 years and was 17% in
66
2012 (9). Of the HIV-infected adults eligible for antiretroviral treatment on the guidelines applicable
67
during the study period (CD4+ T cell count 38ºC) or patient-reported fever, (2) cough or sore throat, and (3)
85
shortness of breath, or difficulty breathing. Only patients with symptom duration of 7 days or less
86
were included in this analysis. Admission criteria and in-hospital treatment protocols (including
87
intensive care (ICU) admission) were at the discretion of the attending physician. There is limited
88
access to ICU beds in South African public hospitals, protocols for ICU admission vary between
89
hospitals and are based on the number of beds available. Oxygen therapy is however, widely available
90
in the general ward setting. All patients admitted (those who spent a night in hospital) from Monday
91
through Friday were eligible for enrolment and enrolled if the WHO case definition was met. At
92
CHBAH, due to the large numbers of subjects and limited resources, enrolment was limited to two of
93
every five working days per week; the selected days varied systematically. Logs were maintained for
94
the total numbers of hospitalized individuals (including those enrolled over weekends and public
95
holidays) who met the study case definitions and of those enrolled by the study-staff. Structured case
96
investigation forms were completed for demographic and clinical information. Enrolled subjects were
97
followed up until death, out-referral to a step-down facility or discharge. A nasopharyngeal (NP) and
98
an oropharyngeal (OP) swab were collected for respiratory virus testing. Blood samples were
99
collected for pneumococcal lytA polymerase chain reaction (PCR) testing and anonymous linked HIV
100
testing by HIV-ELISA was undertaken for those subjects among whom HIV-testing was not
101
undertaken by the attending physician. The collection of blood for bacterial culture and sputum for
102
Mycobacterium tuberculosis (TB) testing was done at the discretion of the attending physicians.
SC
M AN U
TE D
EP
AC C
103
RI PT
81
104
Laboratory methods
105
Upper respiratory secretion samples (combined NP and OP) were transported in a single vial of viral
106
transport medium at 4-8°C to reach the National Institute for Communicable Diseases (NICD) within
107
72 hours of collection. Any storage of sample at the surveillance site was in a temperature regulated
108
refrigerator (4-8°C). Respiratory specimens were tested by multiplex real-time reverse-transcription 5
ACCEPTED MANUSCRIPT 109
PCR (RT PCR) assay for 10 respiratory viruses (respiratory syncytial virus; influenza A and B
110
viruses, parainfluenza virus 1, 2 and 3; enterovirus; human metapneumovirus; adenovirus and
111
rhinovirus) as detailed elsewhere (12). Streptococcus pneumoniae was identified by quantitative real-
112
time PCR detecting the lytA gene from whole blood specimens (13).
RI PT
113 Evaluation of HIV infection status
115
HIV testing was requested by admitting physicians when clinically indicated according to standard
116
practice. This included HIV enzyme-linked immunosorbent assay (ELISA) testing with confirmation
117
by ELISA on a second specimen. In addition, for consenting patients, an anonymized linked dried
118
blood spot specimen was taken for determination of HIV status by ELISA.
SC
114
M AN U
119 Definitions
121
Underlying chronic medical conditions included asthma, chronic lung disease, chronic heart disease,
122
liver disease, renal disease, diabetes mellitus, immune-compromising conditions other than HIV
123
infection based on solicitation from the individual and medical-record review.
124
TE D
120
Calculation of incidence
126
The incidence of RSV-associated SARI hospitalizations was estimated for Soweto (CHBAH) – the
127
only site for which population denominator data were available. CHBAH is the only public hospital
128
serving a community of about 1 million (persons aged ≥ 18 years in 2012). We estimated the
129
incidence of RSV-associated SARI hospitalizations per 100,000 population, the numerator was
130
calculated using the number of RSV-positive SARI hospitalizations and adjusting for non-enrolment
131
(refusal to participate, non-enrolment during weekends and limited sampling of two of the five adult
132
wards) and the dominator being the mid-year population for each of the age groups. All individuals
133
hospitalized with SARI during the study period were logged and this was used to estimate the number
134
of SARI cases not enrolled during the study period. The actual number of RSV-associated cases was
135
then multiplied by the adjustment factor. Adjustment factors were calculated for each age group (18-
136
44 years, 45-64 years and ≥ 65 years). Incidence calculations using the adjusted number of cases,
AC C
EP
125
6
ACCEPTED MANUSCRIPT stratified by age group and HIV status and divided by the mid-year age specific population estimates
138
for each year (14). The HIV prevalence in the study population, by age group, was estimated from the
139
projections of the Actuarial Society of South Africa (ASSA) AIDS and Demographic model (9) and
140
used to impute denominators for numbers of HIV-infected individuals in the population. Confidence
141
intervals for incidence estimates were calculated using the Poisson distribution. Age-specific and
142
overall age-adjusted incidence of RSV-associated SARI in HIV-infected and HIV-uninfected persons
143
was determined using log-binomial regression.
144
(surveillance started late in 2009) of surveillance 2010, 2011 and 2012. In 2013 enrolment was down-
145
scaled at this site and so incidence calculations were not done for this year.
146
Seasonality
147
A description of the seasonality of RSV-associated SARI and total SARI hospitalizations was done
148
by calculating the detection rate of RSV (number of cases of RSV divided by total SARI cases) by
149
month for each year of the study.
RI PT
137
150
M AN U
SC
Incidence is calculated for the complete years
Statistical analysis
152
A descriptive analysis was conducted for demographic and clinical factors associated with SARI in
153
those individuals with RSV-associated SARI. A separate univariate and multivariable analysis by
154
logistic regression was conducted to compare HIV-infected and HIV-uninfected patients with RSV-
155
associated SARI, in which all available information on cases were used. A sensitivity analysis was
156
conducted to explore the effect of missing HIV results on our findings. We repeated the analysis in
157
two scenarios, first considering all those with missing result as HIV-uninfected and secondly
158
considering all those with missing results as HIV-infected. Age group, duration of hospitalization,
159
duration of symptoms and year were defined as categorical variables in multiple levels. All other
160
variables were defined as the presence or absence of the attribute excluding missing data. Multivariate
161
models were explored for univariate association with p value of
S0163-4453(17)30225-6
DOI:
10.1016/j.jinf.2017.06.007
Reference:
YJINF 3947
To appear in:
Journal of Infection
Received Date: 23 April 2017 Revised Date:
22 June 2017
Accepted Date: 24 June 2017
Please cite this article as: Moyes J, Walaza S, Pretorius M, Groome M, von Gottberg A, Wolter N, Haffejee S, Variava E, Cohen AL, Tempia S, Kahn K, Dawood H, Venter M, Cohen C, Madhi SA, South African Severe Acute Respiratory Illness (SARI) Surveillance Group, Respiratory syncytial virus in adults with severe acute respiratory illness in a high HIV prevalence setting, Journal of Infection (2017), doi: 10.1016/j.jinf.2017.06.007. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Respiratory syncytial virus in adults with severe acute respiratory illness in a high HIV prevalence setting. Authors: Jocelyn Moyes1,2, Sibongile Walaza1,
2
, Marthi Pretorius3 , Michelle Groome4, Anne von
RI PT
Gottberg1, Nicole Wolter1, Sumayya Haffejee5, Ebrahim Variava6, Adam L Cohen7, Stefano Tempia1,8 , Kathleen Kahn9 , Halima Dawood10 , Marietjie Venter3, Cheryl Cohen1,2 Shabir A Madhi1,4for the South African Severe Acute Respiratory Illness (SARI) Surveillance Group 1
Centre for Respiratory Diseases and Meningitis, National Institute for
SC
Affiliations:
M AN U
Communicable Diseases of the National Health Laboratory Service, Johannesburg, South Africa; 2School of Public Health, Health Sciences Faculty, University of the Witwatersrand, Johannesburg, South Africa. 3Department of Medical Virology, University of Pretoria, 4
Medical Research Council: Respiratory and Meningeal Pathogens Research Unit and
Department of Science and Technology/National Research Foundation: Vaccine Preventable
TE D
Diseases, University of the Witwatersrand; Johannesburg; South Africa;
5
National health
Laboratory Service; 6Klerksdorp-Tshepong Hospital Complex and The University of the
EP
Witwatersrand; 7Strategic Information Group, Expanded Programme on Immunization, Department of Immunization, Vaccines and Biologicals, World Health Organization
8
AC C
Influenza Program, Centers for Disease Control and Prevention, Pretoria, South Africa and Influenza Division, Centers for Disease Control and Prevention, Atlanta, Georgia, United States of America;
9
MRC/Wits Rural Public Health and Health Transitions Research Unit
(Agincourt), School of Public Health, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa Centre for Global Health Research and Umeå University, Umeå, Sweden and INDEPTH network Accra Ghana;
10
Pietermaritzburg
Hospital Complex and University of KwaZulu-Natal, Pietermaritzburg, South Africa;+
ACCEPTED MANUSCRIPT Corresponding author: Jocelyn Moyes, Centre for Respiratory Diseases and Meningitis, National Institute for Communicable Diseases, Private Bag X4, Sandringham, 2131, Gauteng, South Africa, Telephone: 27 11 386 6410. Fax: 27 11 882 9979. E-mail:
AC C
EP
TE D
M AN U
SC
Word count: Text 3206 words, Abstract 197
RI PT
[email protected]
ACCEPTED MANUSCRIPT 1
Respiratory syncytial virus in adults with severe acute respiratory illness in a high HIV
2
prevalence setting.
3 Authors:
5
Jocelyn Moyes1,2, Sibongile Walaza1, 2 , Marthi Pretorius3 , Michelle Groome4, Anne von Gottberg1,
6
Nicole Wolter1, Sumayya Haffejee5, Ebrahim Variava6, Adam L Cohen7, Stefano Tempia1,8 , Kathleen
7
Kahn9 , Halima Dawood10 , Marietjie Venter3, Cheryl Cohen1,2 Shabir A Madhi1,4for the South African
8
Severe Acute Respiratory Illness (SARI) Surveillance Group
9
Affiliations: 1Centre for Respiratory Diseases and Meningitis, National Institute for Communicable
10
Diseases of the National Health Laboratory Service, Johannesburg, South Africa; 2School of Public
11
Health, Health Sciences Faculty, University of the Witwatersrand, Johannesburg, South Africa.
12
3
13
and Meningeal Pathogens Research Unit and Department of Science and Technology/National
14
Research Foundation: Vaccine Preventable Diseases, University of the Witwatersrand; Johannesburg;
15
South Africa; 5 National health Laboratory Service; 6Klerksdorp-Tshepong Hospital Complex and The
16
University of the Witwatersrand;
17
Immunization, Department of Immunization, Vaccines and Biologicals, World Health Organization
18
Influenza Program, Centers for Disease Control and Prevention, Pretoria, South Africa and Influenza
19
Division, Centers for Disease Control and Prevention, Atlanta, Georgia, United States of America;
20
MRC/Wits Rural Public Health and Health Transitions Research Unit (Agincourt), School of Public
21
Health, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa
22
Centre for Global Health Research and Umeå University, Umeå, Sweden and INDEPTH network
23
Accra Ghana;
24
Pietermaritzburg, South Africa;+
25
Corresponding author: Jocelyn Moyes, Centre for Respiratory Diseases and Meningitis, National
26
Institute for Communicable Diseases, Private Bag X4, Sandringham, 2131, Gauteng, South Africa,
27
Telephone: 27 11 386 6410. Fax: 27 11 882 9979. E-mail: [email protected]
M AN U
SC
RI PT
4
TE D
Department of Medical Virology, University of Pretoria, 4Medical Research Council: Respiratory
Strategic Information Group, Expanded Programme on 8
9
AC C
EP
7
10
Pietermaritzburg Hospital Complex and University of KwaZulu-Natal,
1
ACCEPTED MANUSCRIPT 28
29
Word count: Text 3479 words, Abstract 197
30
AC C
EP
TE D
M AN U
SC
RI PT
31
2
ACCEPTED MANUSCRIPT Abstract
33
Background
34
There are limited data on the epidemiology of respiratory syncytial virus (RSV) illness in HIV-
35
infected adults or the elderly in Africa. We studied the epidemiology of RSV-associated severe acute
36
respiratory illness (SARI) hospitalizations in adults in South Africa from 2009 through 2013.
37
Methods
38
Individuals admitted to sentinel surveillance hospitals were investigated by respiratory tract swabs for
39
RSV, using a multiplex real-time polymerase chain reaction assay. The incidence of RSV-associated
40
SARI was calculated for the one site with population denominators.
41
Results
42
Of 7796 participants investigated, 329 (4%) tested positive for RSV. On multivariable analysis, HIV-
43
infected individuals with RSV-associated SARI had greater odds of being in the age groups 18-44 and
44
45-64 years (odd ratios (OR) 26.3; 95% confidence interval (CI) 6.2-112.1 and OR 11.4; 95% CI 2.6-
45
50.0) compared with those ≥65 years and being female (OR 2.7; 95% CI 1.4-5.4). The relative risk of
46
hospitalization with RSV-associated SARI was 12 to 18 times higher in HIV infected individual
47
compared to that of HIV-uninfected.
48
Conclusion
49
The incidence of RSV-associated SARI was higher in HIV-infected individuals and those aged 65
50
years and older. Further studies are warranted to describe the disease association of RSV detected in
51
adults with SARI.
SC
M AN U
TE D
EP
AC C
52
RI PT
32
3
ACCEPTED MANUSCRIPT 53 Introduction
55
There are limited data on the epidemiology of respiratory syncytial virus (RSV) infection in adults
56
and older children in high HIV prevalence settings like South Africa. In high income countries with a
57
lower HIV burden, the burden of hospitalization with RSV in the elderly has been described as similar
58
to that of influenza (1,2,3,4). Also, RSV infection has been reported to be more severe in the elderly,
59
people living in long term care facilities, immune compromised individuals (specifically those with
60
stem cell and lung transplants) and individuals receiving chemotherapy (4,5,6). Other risk factors in
61
adults include underlying cardiac disease and chronic respiratory illness (7). In a prospective hospital-
62
based surveillance programme in the United States amongst elderly (≥65 years), RSV was identified
63
in 11% of pneumonia episodes, 11% of chronic obstructive pulmonary disease episodes, 5% of
64
congestive heart failure events and 7% of asthma cases (8).
65
In South Africa the highest HIV prevalence occurs among adults aged 15-49 years and was 17% in
66
2012 (9). Of the HIV-infected adults eligible for antiretroviral treatment on the guidelines applicable
67
during the study period (CD4+ T cell count 38ºC) or patient-reported fever, (2) cough or sore throat, and (3)
85
shortness of breath, or difficulty breathing. Only patients with symptom duration of 7 days or less
86
were included in this analysis. Admission criteria and in-hospital treatment protocols (including
87
intensive care (ICU) admission) were at the discretion of the attending physician. There is limited
88
access to ICU beds in South African public hospitals, protocols for ICU admission vary between
89
hospitals and are based on the number of beds available. Oxygen therapy is however, widely available
90
in the general ward setting. All patients admitted (those who spent a night in hospital) from Monday
91
through Friday were eligible for enrolment and enrolled if the WHO case definition was met. At
92
CHBAH, due to the large numbers of subjects and limited resources, enrolment was limited to two of
93
every five working days per week; the selected days varied systematically. Logs were maintained for
94
the total numbers of hospitalized individuals (including those enrolled over weekends and public
95
holidays) who met the study case definitions and of those enrolled by the study-staff. Structured case
96
investigation forms were completed for demographic and clinical information. Enrolled subjects were
97
followed up until death, out-referral to a step-down facility or discharge. A nasopharyngeal (NP) and
98
an oropharyngeal (OP) swab were collected for respiratory virus testing. Blood samples were
99
collected for pneumococcal lytA polymerase chain reaction (PCR) testing and anonymous linked HIV
100
testing by HIV-ELISA was undertaken for those subjects among whom HIV-testing was not
101
undertaken by the attending physician. The collection of blood for bacterial culture and sputum for
102
Mycobacterium tuberculosis (TB) testing was done at the discretion of the attending physicians.
SC
M AN U
TE D
EP
AC C
103
RI PT
81
104
Laboratory methods
105
Upper respiratory secretion samples (combined NP and OP) were transported in a single vial of viral
106
transport medium at 4-8°C to reach the National Institute for Communicable Diseases (NICD) within
107
72 hours of collection. Any storage of sample at the surveillance site was in a temperature regulated
108
refrigerator (4-8°C). Respiratory specimens were tested by multiplex real-time reverse-transcription 5
ACCEPTED MANUSCRIPT 109
PCR (RT PCR) assay for 10 respiratory viruses (respiratory syncytial virus; influenza A and B
110
viruses, parainfluenza virus 1, 2 and 3; enterovirus; human metapneumovirus; adenovirus and
111
rhinovirus) as detailed elsewhere (12). Streptococcus pneumoniae was identified by quantitative real-
112
time PCR detecting the lytA gene from whole blood specimens (13).
RI PT
113 Evaluation of HIV infection status
115
HIV testing was requested by admitting physicians when clinically indicated according to standard
116
practice. This included HIV enzyme-linked immunosorbent assay (ELISA) testing with confirmation
117
by ELISA on a second specimen. In addition, for consenting patients, an anonymized linked dried
118
blood spot specimen was taken for determination of HIV status by ELISA.
SC
114
M AN U
119 Definitions
121
Underlying chronic medical conditions included asthma, chronic lung disease, chronic heart disease,
122
liver disease, renal disease, diabetes mellitus, immune-compromising conditions other than HIV
123
infection based on solicitation from the individual and medical-record review.
124
TE D
120
Calculation of incidence
126
The incidence of RSV-associated SARI hospitalizations was estimated for Soweto (CHBAH) – the
127
only site for which population denominator data were available. CHBAH is the only public hospital
128
serving a community of about 1 million (persons aged ≥ 18 years in 2012). We estimated the
129
incidence of RSV-associated SARI hospitalizations per 100,000 population, the numerator was
130
calculated using the number of RSV-positive SARI hospitalizations and adjusting for non-enrolment
131
(refusal to participate, non-enrolment during weekends and limited sampling of two of the five adult
132
wards) and the dominator being the mid-year population for each of the age groups. All individuals
133
hospitalized with SARI during the study period were logged and this was used to estimate the number
134
of SARI cases not enrolled during the study period. The actual number of RSV-associated cases was
135
then multiplied by the adjustment factor. Adjustment factors were calculated for each age group (18-
136
44 years, 45-64 years and ≥ 65 years). Incidence calculations using the adjusted number of cases,
AC C
EP
125
6
ACCEPTED MANUSCRIPT stratified by age group and HIV status and divided by the mid-year age specific population estimates
138
for each year (14). The HIV prevalence in the study population, by age group, was estimated from the
139
projections of the Actuarial Society of South Africa (ASSA) AIDS and Demographic model (9) and
140
used to impute denominators for numbers of HIV-infected individuals in the population. Confidence
141
intervals for incidence estimates were calculated using the Poisson distribution. Age-specific and
142
overall age-adjusted incidence of RSV-associated SARI in HIV-infected and HIV-uninfected persons
143
was determined using log-binomial regression.
144
(surveillance started late in 2009) of surveillance 2010, 2011 and 2012. In 2013 enrolment was down-
145
scaled at this site and so incidence calculations were not done for this year.
146
Seasonality
147
A description of the seasonality of RSV-associated SARI and total SARI hospitalizations was done
148
by calculating the detection rate of RSV (number of cases of RSV divided by total SARI cases) by
149
month for each year of the study.
RI PT
137
150
M AN U
SC
Incidence is calculated for the complete years
Statistical analysis
152
A descriptive analysis was conducted for demographic and clinical factors associated with SARI in
153
those individuals with RSV-associated SARI. A separate univariate and multivariable analysis by
154
logistic regression was conducted to compare HIV-infected and HIV-uninfected patients with RSV-
155
associated SARI, in which all available information on cases were used. A sensitivity analysis was
156
conducted to explore the effect of missing HIV results on our findings. We repeated the analysis in
157
two scenarios, first considering all those with missing result as HIV-uninfected and secondly
158
considering all those with missing results as HIV-infected. Age group, duration of hospitalization,
159
duration of symptoms and year were defined as categorical variables in multiple levels. All other
160
variables were defined as the presence or absence of the attribute excluding missing data. Multivariate
161
models were explored for univariate association with p value of
