Epidemiological studies on tuberculosis control and respiratory viruses

(1)

Sloot, R.

Publication date

2015

Document Version

Final published version

Link to publication

Citation for published version (APA):

Sloot, R. (2015). Epidemiological studies on tuberculosis control and respiratory viruses.

General rights

It is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), other than for strictly personal, individual use, unless the work is under an open content license (like Creative Commons).

Disclaimer/Complaints regulations

If you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, stating your reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Ask the Library: https://uba.uva.nl/en/contact, or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam, The Netherlands. You will be contacted as soon as possible.

(2)

Chapter

6

Distribution and viral load of

respiratory viruses differ by illness

severity in adults:

a comparison between

community and

hospital populations

Rosa Sloot Maarten F. Schim van der Loeff Richard Molenkamp Hetty W.M. van Eijk Sjoerd P.H. Rebers Marieke B. Snijder Maria Prins Janke Schinkel Menno D. de Jong In preparation

(3)

Abstract

Objective: To better understand the clinical significance of respiratory viruses, we investigated the prevalence, relative distribution and viral load of respiratory viruses in community and hospital populations by illness severity.

Methods: From 2011 to 2013, nasopharyngeal samples were collected during the influenza season from adult participants of the HELIUS study, a Dutch population-based, multi-ethnic cohort study. For comparison, routine diagnostic and demographic data were used of adult patients presenting at a hospital serving the geographical area of the cohort study. Study participants were grouped by approximated illness severity: asymptomatic and mildly symptomatic HELIUS participants and hospital outpatients, inpatients, and ICU-admitted patients. Respiratory viruses were detected by multiplex real-time PCR. Crossing point values were used to estimate viral load. Results: Respiratory viruses were detected in 12% of the community population and 31% of the hospital population. Among virus-positive subjects, rhinovirus (RV), human coronavirus (hCoV) and human bocavirus were significantly overrepresented in the community population, while the reverse was true for influenza A virus (InfA) and human metapneumovirus. Viral load of InfA and RSV was significantly correlated with illness severity. Correlations were less clear for RV and hCoV but highest viral loads were observed in ICU-admitted patients.

Conclusion: Differences in distribution between community and hospital populations confirm differences in pathogenicity between respiratory viruses in adults. Viral load correlated with illness severity for InfA and RSV but this was less clear for viruses with reduced pathogenicity, like RV and hCoV. Determining the clinical significance of such viruses in individual hospitalized patients remains challenging.

(4)

6 Introduction

Rapid and accurate detection of viral pathogens facilitates early diagnosis, early identification of outbreaks, timely intervention, effective management of high-risk contacts, appropriate antimicrobial therapy, and avoidance of unnecessary laboratory testing [1]. The introduction of real-time polymerase chain reaction (RT-PCR)-based methods has greatly improved the diagnostics for respiratory viral infections. However, detection of a viral pathogen by RT-PCR does not necessarily imply causality of the illness; it is well established that subclinical respiratory viral infections do occur. Moreover, because of its high analytic sensitivity, RT-PCR may detect small amounts of viral nucleic acids, the clinical relevance of which may be difficult to interpret since, for example, these may reflect past or asymptomatic infections [2]. Multiple studies have demonstrated a correlation between viral load of respiratory viruses and severe disease [3] [4] [5] [6] [7]. Hence, quantitation of viral nucleic acids may be helpful for the clinical interpretation of a positive PCR result.

Understanding the clinical significance of respiratory viral infections is essential for improving preventive and therapeutic strategies. Most etiological studies have focused on patients presenting in health care settings with respiratory illness. However, studies in the general population could provide information on the background prevalence of respiratory viral infections and thereby contribute to our understanding of the clinical interpretation of a positive PCR result in patients with respiratory illness who are seeking health care. Therefore, the objective of our study was to compare the prevalence and relative distribution of respiratory viral infections among adult populations that vary in illness severity, ranging from asymptomatic or mildly symptomatic individuals in the general population not seeking healthcare to patients with severe disease requiring intensive care, and to evaluate whether viral load estimations could aid in the interpretation of diagnostic test results. A population-based cohort study provided the opportunity to study the prevalence, relative distribution and viral load of respiratory viral infections in the adult general population. For comparison, demographic data and routine diagnostic test results were used from adult patients who, during the same period as the cohort study, were presented or admitted to a large tertiary referral hospital which also provides non-tertiary care for the catchment area of the cohort study population.

(5)

Methods

Study population

HELIUS study

The HELIUS study (acronym for Healthy Life in an Urban Setting) is designed as a large prospective cohort study, initiated by the Academic Medical Center (AMC) and the Public Health Service of Amsterdam, to understand the unequal burden of disease across ethnic groups [8]. People aged 18-70 years, of Dutch, Turkish, Moroccan, Ghanaian, Surinamese origin were sampled from the municipal civil register of Amsterdam and invited to participate. Sampling was random but stratified by ethnicity, as defined by the country of birth of individuals or their parents as documented in the civil register [8]. Persons who were unable to give informed consent and persons who were not registered with a general practitioner were excluded. Baseline data collection started in January 2011 and is still ongoing. Data are collected through a questionnaire and physical examination, and biological samples are obtained during study visits. During the data collection from January 2011 until June 2013, nasal and throat swabs were obtained by trained nurses or research assistants using flocked swabs (FLOQSwabsTM_{, Copan, Brescia, Italy) and collected in a single tube containing 3 mL of}

viral transport medium (UTM™ medium, Copan, Brescia, Italy). The transport medium was kept at room temperature until transport to the AMC Department of Medical Microbiology on the same day. On arrival, the tubes containing the combined swabs and transport medium were vortexed, the swabs were discarded, and the medium was divided into two equal aliquots and stored at -80°C until further processing [8]. Before the physical examination at the study site, participants were asked by a trained research assistant, if they currently experienced any of the following seven symptoms: fever, headache, muscular pain, cough, a sore throat, shortness of breath, or a runny nose, or had experienced these symptoms in the 2 weeks preceding the day of the examination.

At the time of study initiation few participants of Turkish and Moroccan origin had been recruited, and therefore only participants from three ethnic groups (Dutch, Surinamese and Ghanaian) were selected for inclusion in the current study. Of participants recruited during the influenza seasons (October – March) in the years

(6)

6

2011 until and including March 2013, 100 symptomatic and 100 asymptomatic participants were randomly selected from each ethnic group. Participants were defined as ‘symptomatic’ if they reported at least either a runny nose or both fever and cough on the day of sample collection, irrespective of the presence of symptoms in the 2 weeks preceding sample collection. Participants were defined as ‘asymptomatic’ if they had none of the seven symptoms on the day of sample collection or in the 2 weeks preceding that day.

Patients presenting at the hospital

At the AMC, a tertiary referral centre which also provides non-tertiary care for the local area, nasopharyngeal samples are routinely collected and analysed in real time from patients clinically suspected of respiratory tract infections who are presenting at outpatient clinics or are admitted to the hospital. Types of flocked swabs and collection medium as well as the multiplex RT-PCR platform used for these routine diagnostics are identical to those used in the HELIUS population. Demographic variables and RT-PCR test results are registered in electronic registration and laboratory systems. Anonymized demographic data and diagnostic results of patients aged ≥18 years with a clinical suspicion of respiratory tract infection, and sampled during the influenza seasons of the years 2011, 2012 and 2013, were included in the analysis.

Virological assay

Extraction of nucleic acids from 200 µl of viral transport medium was performed by MagNA Pure extraction using the total nucleic acid extraction kit (Roche Diagnostics, Penzberg, Germany). The presence of viral pathogens in samples was detected by multiplex reverse transcriptase RT-PCR as previously described [5]. Respiratory viruses detected in this assay included: rhinovirus (RV; A, B, and C), human coronavirus (hCoV: HKU1, NL63, 229E and OC43), influenza A virus (InfA), respiratory syncytial virus (RSV; A and B), influenza B virus (InfB), enterovirus (EV), adenovirus (AdV), human metapneumovirus (hMPV), parainfluenza viruses (PIVs 1, 2, 3, 4), human bocavirus (hBoV) and parechovirus (PeV).

The crossing point (C_p) value, reflecting the cycle at which a positive PCR signal is detected, was calculated using LC480 software (Roche Diagnostics, Penzberg, Germany). It was used as an approximation of the amount of viral nucleic acids

(7)

present, as C_p values are inversely correlated with the number of target DNA or RNA molecules present in the sample (in this study denoted as viral load). A difference in C_p values of 3 approximately represents a 10-fold difference in the levels of target nucleic acids. A sample was considered positive if it passed the internal positive controls and if C_p value ≤40, and negative if the value was above 40.

Statistical analysis

The study population was divided in groups according to symptom status and health care use, which was used as an approximation of illness severity, ranging from (1) asymptomatic and (2) mildly symptomatic HELIUS participants, reflective of the general population who did not seek hospital care, to three groups of symptomatic patients seeking hospital care: (3) patients presenting at general outpatient clinics, (4) patients admitted to the hospital (but not to intensive care units [ICUs]), and (5) patients admitted to ICUs. Patients presenting at the Emergency Department were excluded due to the expected high variability in severity at time of presentation within this group of patients.

Analyses of the complete study population were done to compare characteristics of individuals and prevalence of respiratory viral infections between groups (Table 1), and to assess the association between PCR-positivity and approximated illness severity based on symptom status and health care use (Table S1). In addition, among individuals with positive PCR results, we compared the distribution of viral species between groups (Table 2), and analysed possible relationships between viral load and approximated illness severity (Table 3, Figure 1, Table S2). Analyses are described in more detail below.

Differences in categorical variables were assessed using the Chi-squared test or Fisher exact test; for continuous variables Mann-Whitney U test or Kruskal-Wallis were used. Three comparisons were made in order to assess differences (1) between the symptomatic and asymptomatic HELIUS groups reflecting the general population not seeking hospital care, (2) between the three groups presenting at or admitted to the hospital, and (3) between all five groups.

Associations between groups and PCR positivity were assessed using logistic regression analysis. Because the number of positive PCR results among HELIUS

(8)

6

participants was small, the virus specific comparisons were restricted to the four most commonly detected viruses in the HELIUS population: RV, hCoV, InfA, and RSV. Sex, age, month and year of sampling were a priori included in the logistic regression models.

To investigate the relative distribution of viruses between groups in PCR-positive individuals, the two groups of HELIUS participants were combined (groups 1, 2) and were compared to each of the hospital groups separately (groups 3, 4 and 5) as well as combined (group 1/2 vs group 3-5). Comparisons between groups were made for each respiratory virus detected by multiplex RT-PCR.

Spearman’s rank-order correlation test and linear regression analysis were done to determine relationships between viral load of RV, hCoV, InfA, RSV and approximated illness severity of groups. Both analyses were done with all groups as well as with inclusion of hospital groups only (group 3-5). Sex, age, month and year of sampling were a priori included in the linear regression models.

The level of significance in all analyses was P<0.05 and analyses were done in SPSS 22.0 (SPSS, Chicago, USA).

Ethics

The HELIUS study was approved by the AMC Medical Ethics Committee and informed consent was obtained from each participant at study entry. According to the Dutch law on medical research (WMO), article 1, no ethical approval is required when using anonymous data from routine diagnostic databases, as was done for the diagnostic and demographic data of AMC patients. The study was conducted according to the Dutch code of conduct for responsible use of human tissue and medical research 2011 [9].

Results

Study population

Among 3,417 HELIUS participants enrolled during the period of study, 600 adult participants were selected for inclusion in the current study. From 588 of these,

(9)

respiratory samples were available. During the same study period, respiratory samples of 600 adult patients presenting or admitted to the hospital with clinically suspected respiratory tract infection were collected and analysed. In Table 1 characteristics of the groups are reported.

The 291 asymptomatic and 297 symptomatic HELIUS participants were comparable with regard to ethnicity, education, and smoking status, but asymptomatic participants were more often male and were older. The hospital groups were comparable with regard to sex but there were significant differences between groups with regard to age (Table 1). ICU patients were older than outpatients and inpatients (both, P<0.001). HELIUS and hospital groups were not comparable with regard to gender and age (Table 1). HELIUS participants were younger (median age 49, interquartile range [IQR]=35-56) than hospital patients (median age 56, IQR=39-67) (P<0.001). Among HELIUS participants, at least one viral pathogen was detected in 12% of participants, with a higher detection rate among symptomatic as compared to asymptomatic participants (18% versus 6%, P<0.001) (Table 1). The most prevalent viruses detected were RV (5%) and hCoV (3%), both of which were found significantly more often in symptomatic individuals. Influenza viruses were only detected in symptomatic and not in asymptomatic individuals. In 31% of patients presenting at or admitted to the hospital at least one viral pathogen was detected, and the most prevalent viruses were InfA (8%) and RV (8%). Comparisons of viral prevalence across five groups revealed significant differences for RV, InfA, RSV, InfB, hMPV, but not for the remaining viruses. Multivariable logistic regression analysis showed that the odds of viral prevalence across groups remained significantly different for RV, InfA and RSV (InfB, hMPV were not investigated) (Table S1).

There were no associations between demographic and epidemiological characteristics and PCR positivity for any respiratory virus, investigated in multivariable logistic regression analysis for each of the five groups separately (data not shown).

(10)

6

Table 1. C har act eristics o f the stud y population b y appr oximat ed illness se verity Asympt omatic HELIUS Sympt omatic HELIUS P-value 1 Outpatients Inpatients ICU patients P-value 2 P-value 3 n (%)* n (%)* n (%)* n (%)* n (%)* Total 291 297 116 284 200 Se x M ale 155 (53%) 120 (40%) 0.002 63 (54%) 149 (52%) 103 (52%) 0.890 0.008 Female 136 (47%) 177 (60%) 53 (46%) 135 (48%) 97 (49%) Age (y ears) M edian (IQR) 49 (40-56) 48 (32-55) 0.014 4 53 (38-63) 52 (36-65) 63 (50-74) <0.001 5 <0.001 5 18-34 56 (19%) 81 (27%) 0.039 25 (22%) 67 (24%) 21 (11%) <0.001 <0.001 35-54 135 (46%) 135 (46%) 37 (32%) 85 (30%) 43 (22%) ≥55 100 (34%) 81 (27%) 54 (47%) 132 (47%) 136 (68%) Ethnicity Dut ch 96 (33%) 99 (33%) 0.995 Surinamese 96 (33%) 98 (33%) Ghanaian 99 (34%) 100 (34%) Educ ation (le ve l 6) 1 33 (11%) 40 (14%) 0.602 2 89 (31%) 75 (25%) 3 78 (27%) 79 (27%) 4 90 (31%) 83 (28%) Unkno wn 1 (0.3%) 20 (7%) Smoking N eve r 174 (60%) 157 (53%) 0.630 Former smok er 62 (21%) 60 (20%) Curr ent 54 (19%) 60 (20%) Unkno wn 1 (0.3%) 20 (7%)

(11)

Table 1 (c ontinue d). C har act eristics o f the stud y population b y appr oximat ed illness se verity Asympt omatic HELIUS Sympt omatic HELIUS P-value 1 Outpatients Inpatients ICU patients P-value 2 P-value 3 n (%)* n (%)* n (%)* n (%)* n (%)* Comor bidities 7 No 169 (58%) 120 (40%) 0.001 Yes (CDC risk f act or f or r esp . in f.) 86 (30%) 102 (34%) Yes (other) 36 (12%) 56 (19%) Unkno wn 0 (0%) 19 (6%) PCR positi

ve For at least one vir

us 16 (6%) 54 (18%) <0.001 41 (35%) 80 (28%) 66 (33%) 0.294 <0.001 Rhino vir us 6 (2%) 22 (7%) 0.002 13 (11%) 23 (8%) 13 (7%) 0.338 0.004 H uman c or ona vir us 5 (2%) 15 (5%) 0.026 8 (7%) 12 (4%) 12 (6%) 0.490 0.082 In fluenza A vir us 0 (0%) 6 (2%) 0.031 4 (3%) 26 (9%) 20 (10%) 0.100 <0.001 Respir at or y syncytial vir us 2 (1%) 4 (1%) 0.686 8 (7%) 11 (4%) 6 (3%) 0.234 0.003 In fluenza B vir us 0 (0%) 1 (0.3%) 1.000 4 (3%) 4 (1%) 6 (3%) 0.346 0.001 Ent er ovir us 0 (0%) 1 (0.3%) 1.000 1 (1%) 0 (0%) 0 (0%) 0.193 0.214 Adeno vir us 0 (0%) 1 (0.3%) 1.000 1 (1%) 1 (0.4%) 2 (1%) 0.535 0.312 H uman M etapneumo vir us 0 (0%) 1 (0.3%) 1.000 6 (5%) 6 (2%) 8 (4%) 0.246 <0.001 Par ain fluenza vir us 1 (0.3%) 3 (1%) 0.624 2 (2%) 0 (0%) 1 (1%) 0.081 0.181 H uman boc avir us 1 (0.3%) 4 (1%) 0.373 0 (0%) 1 (0.4%) 1 (1%) 1.000 0.578 Par ec ho vir us 1 (0.3%) 1 (0.3%) 1.000 0 (0%) 0 (0%) 0 (0%) NA 1.000 M onth o f sampling O ct ober 54 (19%) 50 (17%) 0.983 10 (9%) 28 (10%) 18 (9%) 0.002 <0.001 N ov ember 42 (14%) 41 (14%) 12 (10%) 26 (9%) 25 (13%) D ec ember 33 (11%) 37 (13%) 11 (10%) 20 (7%) 43 (22%) Januar y 52 (18%) 58 (20%) 35 (30%) 96 (34%) 48 (24%) Fe br uar y 56 (19%) 58 (20%) 25 (22%) 58 (20%) 37 (19%) Ma rc h 54 (19%) 53 (18%) 23 (20%) 56 (20%) 29 (15%)

(12)

6

Table 1 (c ontinue d). C har act eristics o f the stud y population b y appr oximat ed illness se verity Asympt omatic HELIUS Sympt omatic HELIUS P-value 1 Outpatients Inpatients ICU patients P-value 2 P-value 3 n (%)* n (%)* n (%)* n (%)* n (%)* Year o f sampling 2011 30 (10%) 42 (14%) 0.124 55 (47%) 127 (45%) 53 (27%) <0.001 <0.001 2012 158 (54%) 138 (47%) 34 (29%) 88 (31%) 64 (32%) 2013 103 (35%) 117 (39%) 27 (23%) 69 (24%) 83 (42%) * U nless stat ed other wise Unkno wn c at eg ories in italics w er e ex cluded fr om anal ysis . Ethnicity , educ ation, smoking and c omorbidities w er e not r ec or ded f or AMC patients . AMC= A cademic M edic al C ent er; CDC = C ent ers f or Disease C ontr ol and P re vention; HELIUS= H ealth y Li fe I n an U rban S etting; ICU = I nt ensi ve c ar

e unit; IQR = int

er quar tile r ang e; PCR = pol ymer ase c hain r eaction 1 Comparison betw een asympt

omatic and sympt

omatic HELIUS par

tici pants 2 Comparison acr oss the thr ee AMC gr oups (outpatients , I npatients , ICU) 3 Comparison acr

oss all fiv

e gr oups Chi-squar ed t est or Fisher ex act t est w er e used f or c omparisons , unless stat ed other wise . 4 M ann-W hitne y U -test 5 Krusk al-W allis t est 6 Le ve ls w er e c oded as: N ev er been t o sc hool or e lementar y sc hooling onl y 1 Lo w er v oc ational or lo w er sec ondar y sc hooling 2 Int ermediat e v oc ational or int ermediat e/higher sec ondar y sc hooling 3 Higher v oc ational sc hooling or uni versity 4 7 Comorbidities kno wn t o be associat ed with r espir at or y in fections [r ef], and r epor ted b y HELIUS par tici

pants as being diagnosed b

y a ph ysician: Diabet es , asthma, neur ologic al c onditions , lung disease , hear t disease , blood disor ders , li ver disor ders , metabolic disor ders , w eak

ened immune syst

em, and obesity . Other c omorbidities repor ted b y HELIUS par tici

pants as being diagnosed

by a ph ysician inc luded: chr onic f atigue , headac he or migr aine , psoriasis , ecz ema, se ver e bo w el disor ders , ar thr osis , hernia, inc ontinenc e, pain in bod y par ts (bac k, nec k, shoulder , e lbo w, hand, wrist) and c hr

onic muscular pain.

[ref] http://www

.cdc

.g

ov/flu/about/disease/high_risk.

(13)

Table 2. R elati ve distribution o f r espir at or y vir uses in PCR -positi ve indi viduals HELIUS (asympt omatic and sympt omatic) n (%) Outpatients n (%) Inpatients n (%) ICU patients n (%) AMC patients -total- n (%) p-value 1 p-value 2 p-value 3 p-value 4 Total 75 47 84 69 200 Rhino vir us 28 (37%) 13 (28%) 23 (27%) 13 (19%) 49 (25%) 0.271 0.180 0.014 0.035 H uman c or ona vir us 20 (27%) 8 (17%) 12 (14%) 12 (17%) 32 (16%) 0.218 0.052 0.181 0.044 In fluenza A vir us 6 (8%) 4 (9%) 26 (31%) 20 (29%) 50 (25%) 1.000 <0.001 0.001 0.002 Respir at or y syncytial vir us 6 (8%) 8 (17%) 11 (13%) 6 (9%) 25 (13%) 0.128 0.299 0.880 0.293 In fluenza B vir us 1 (1%) 4 (9%) 4 (5%) 6 (9%) 14 (7%) 0.072 0.216 0.055 0.077 Ent er ovir us 1 (1%) 1 (2%) 0 (0%) 0 (0%) 1 (1%) 1.000 0.472 1.000 0.472 Adeno vir us 1 (1%) 1 (2%) 1 (1%) 2 (3%) 4 (2%) 1.000 1.000 0.607 1.000 H uman M etapneumo vir us 1 (1%) 6 (13%) 6 (7%) 8 (12%) 20 (10%) 0.013 0.121 0.014 0.016 Par ain fluenza vir us 4 (5%) 2 (4%) 0 (0%) 1 (1%) 3 (2%) 1.000 0.047 0.368 0.091 H uman B oc avir us 5 (7%) 0 (0%) 1 (1%) 1 (1%) 2 (1%) 0.155 0.101 0.211 0.018 Par ec ho vir us 2 (3%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0.522 0.221 0.497 0.074 1 Comparison betw

een HELIUS par

tici

pants and Outpatients

2 Comparison betw

een HELIUS par

tici

pants and I

npatients

3 Comparison betw

een HELIUS par

tici

pants and ICU patients

4 Comparison betw

een HELIUS par

tici

pants and AMC patients (t

otal) Chi-squar ed t est or Fisher ex act t est w er e used f or c omparisons . AMC= A cademic M edic al C ent er HELIUS= H ealth y Li fe I n an U rban S etting ICU = I nt ensi ve c ar e unit PCR = pol ymer ase c hain r eaction

(14)

6

1.a. Rhinovirus 15 20 25 30 35 40 G-1/2 (n=28) G-3 (n=13) G-4 (n=23) G-5 (n=13)

Correlation between Cp values and groups arranged by approximated illness severity

rs = 0.112, P=0.331 (G1/2-G5) rs = -0.466, P=0.001 (G3-G5) Cp v al ue 1.b. Human coronavirus 15 20 25 30 35 40 45 G-1/2 (n=20) G-3 (n=8) G-4 (n=12) G-5 (n=12)

rs = -0.032, P=0.820 (G1/2-G5) rs = -0.193, P=0.290 (G3-G5) Cp v al ue

(15)

1.c. Influenza A virus 15 20 25 30 35 40 G-1/2 (n=6) G-3 (n=4) G-4 (n=26) G-5 (n=20)

1.d. Respiratory syncytial virus

15 20 25 30 35 40 G-1/2 (n=6) G-3 (n=8) G-4 (n=11) G-5 (n=6)

Figure 1. Box and whisker plots of crossing point values by approximated illness severity of groups The box represents the interquartile range, the horizontal line in the box represents the median, and the whisker top and bottom represent the lowest and highest C_p values (corresponding with the highest and lowest viral

(16)

6

Please see Table 3 for median (IQR) C_p values for each group, by type of virus. AMC= Academic Medical Center

C_p= Crossing point values

HELIUS= Healthy Life In an Urban Setting ICU= Intensive care unit

G-1/2 = HELIUS, asymptomatic and symptomatic G-3 = AMC, outpatients

G-4 = AMC, inpatients G-5 = AMC, ICU patients

Relative proportions of respiratory viruses between groups

When comparing the relative proportions of specific respiratory viruses among PCR-positive individuals in each group, substantial differences were observed (Table 2). While RVs were the most commonly detected viruses in 75 PCR-positive HELIUS participants, InfA represented the largest proportion of viruses detected in 200 positive hospital patients. The proportion of RV was significantly higher among PCR-positive HELIUS participants compared to PCR-PCR-positive hospital patients (37% vs 25%, P=0.035) and was lowest among PCR-positive ICU-admitted inpatients (19%). Similarly, the proportions of hCoV and hBoV were significantly higher in HELIUS participants than in hospital patients (27% vs 16% and 7% vs 1%, respectively (P=0.044 and 0.018)). The opposite was observed for InfA and hMPV, the proportions of which were significantly higher in hospital patients compared to HELIUS participants (25% vs 8% and 10% vs 1%, respectively (P=0.002 and 0.016)) (Table 2). Highest proportions of InfA were observed in hospitalized patients (inpatients 31%, ICU inpatients 29%). Proportions of RSV and InfB were also considerably higher in hospital patients compared to HELIUS participants but these differences did not reach statistical significance.

Relation between viral load and approximated illness severity

Comparisons of viral load estimations between groups were done for the four most commonly detected viruses (RV, hCoV, InfA and RSV) as expressed by C_p values which correlate inversely with viral load levels. For RV and hCoV, no correlations of viral load were observed between the five groups (Table 3, Figure 1). Also in linear regression analysis no significant differences in hCoV load between groups were observed (Table S2). For RV, viral loads were similar between HELIUS participants and ICU-admitted patients, but significantly lower in outpatients and non-ICU

(17)

inpatients (Table S2). When restricted to hospital groups, RV viral loads correlated significantly with approximated severity with highest C_p values (indicating lowest viral load) observed in outpatients and lowest C_p values in ICU-admitted patients (r_s=-0.466, P=0.001) (Table 3, Figure 1). Of note, although no significant correlation was observed, highest viral loads of hCoV were also observed in ICU patients. For InfA, estimated viral loads were lowest among HELIUS participants and increased in the hospital groups (C_p values, r_s=-0.311, P=0.020) (Table 3, Figure 1). While highest InfA viral loads were observed in ICU-admitted patients, no significant correlation of C_p values was observed between hospital groups only. Accordingly, linear regression analyses showed significantly higher InfA viral loads in both groups of hospitalized patients (non-ICU, ICU) when compared to HELIUS participants, but no significant differences between hospital groups (Table S2). Similar to InfA, there was a significant correlation between RSV load and groups, with highest viral load observed in ICU-admitted patients (C_p values, r_s=-0.374, P=0.038) (Table 3, Figure 1). This correlation between RSV load and approximated illness severity was stronger when the analysis was restricted to hospital groups only (C_p values, r_s=-0.441, P=0.027) (Table 3, Figure 1). However, linear regression analysis did not show significant differences of RSV C_p values between groups (Table S2).

(18)

6

Table 3. C rossing point v alues o f positi ve PCR r esults , b y type o f vir us and appr oximat ed illness se verity Gr oup 1 Gr oup 2 Gr oup 1/2 G roup 3 Gr oup 4 Gr oup 5 Gr oup 1/2-5 Gr oup 3-5 Asympt omatic HELIUS Sympt omatic HELIUS Combined HELIUS Outpatients Inpatients ICU patients Vir us Cp , Cp , Cp , Cp , Cp , Cp , rs P-value rs P-value me dian (IQR) me dian (IQR) me dian (IQR) me dian (IQR) me dian (IQR) me dian (IQR) RV 26.9 (25.6-28.4) (n=6) 28.1 (26.9-30.0) (n=22) 27.8 (25.9-29.9) (n=28) 30.8 (29.4-31.6) (n=13) 30.6 29.4-32.0) (n=23) 27.6 (25.0-29.4) (n=13) 0.112 0.331 -0.466 0.001 hC oV 33.6 (28.4-34.9) (n=5) 33.3 (27.8-35.1) (n=15) 33.4 (28.1-34.8) (n=20) 34.8 (30.1-37.1) (n=8) 33.1 (29.3-36.5) (n=12) 30.4 (28.4-36.1) (n=12) -0.032 0.820 -0.193 0.290 Inf A N. A. (n=0) 34.7 (29.1-36.9) (n=6) 34.7 (29.1-36.9) (n=6) 32.5 (31.1-34.0) (n=4) 31.1 (28.7-33.1) (n=26) 30.8 (25.2-32.7) (n=20) -0.311 0.020 -0.205 0.152 RSV N. A. (n=2) 30.7 (27.9-33.4) (n=4) 29.9 (28.6-32.8) (n=6) 33.1 (27.9-35.4) (n=8) 28.6 (26.8-32.7) (n=11) 24.2 (21.9-30.5) (n=6) -0.374 0.038 -0.441 0.027 RV= rhino vir us hC oV= c or ona vir us In fA= in fluenza vir us A RSV= respir at or y syncytial vir us AMC= A cademic M edic al C ent er HELIUS= H ealth y Li fe I n an U rban S etting ICU= I nt ensi ve c ar e unit IQR= int er quar tile r ang e PCR= pol ymer ase c hain r eaction Cp = C rossing point v alues rs = Spearman ’s r ank -or der c orr elation w as done t o det ermine the r elationshi p betw een gr oups arr ang ed b y appr oximat ed illness se verity and C p values in tw o c omparisons: (1) A ll gr oups inc luded (G1/2 thr ough G5) and (2) O nl y the thr ee AMC gr oups inc luded (G3 thr ough G5)

(19)

Discussion

Respiratory viral pathogens are considered important causes of morbidity and mortality. Sensitive RT-PCR assays allow detection of a wide range of respiratory viruses, but determining the clinical significance of a positive PCR result in individual patients remains challenging. In aid of understanding the significance of respiratory viral detection, our population-based cohort (HELIUS) provided the opportunity to study the prevalence, distribution and viral loads of respiratory viral infections in the adult general population and compare these to patients seeking health care in our hospital (AMC), serving the catchment area of the cohort population.

Whilst a large body of literature exists on the prevalence of respiratory viruses in patients seeking health care for acute respiratory illness, prevalence data in the general population are scarce, in particular for adults. In our adult population cohort, respiratory viruses, mostly RV and hCoV, were detected in 18% of mildly symptomatic individuals and 6% of asymptomatic persons. The latter percentage seems in accordance with a previous study in the elderly [10], but reported prevalences of respiratory viral infections in asymptomatic children are substantially higher [5] [11] [12] [13] [14].

The observed significant overrepresentation among detected viruses of RV, hCoV and hBoV in the general population, and of InfA and hMPV in the hospital population confirmed differences in pathogenicity between these respiratory viruses. Strikingly, hMPV represented 12% of detected viruses in ICU-admitted patients, which supports its clinical significance, not only for children [15] [16] [17] but also for the adult population. RSV and InfB were also detected more commonly in hospital patients than in the general population (RSV 13% vs 8% of virus-positives, InfB 7% vs 1%, respectively) but these differences did not reach statistical significance. Combined, the latter four respiratory viruses, for which little controversy exists regarding their pathogenicity, contributed to 55% of detected viruses in the hospital population versus only 19% in the general population. These observations provide reassurance that causality can be implied if these viruses are detected in individual hospitalized patients presenting with acute respiratory illnesses. However, this is much less clear for viruses such as RV and hCoV, which together represented 41% of detected viruses in the hospital population (RV 25%, hCoV 16%). The observed association with symptomatology in the population cohort confirmed a role of both viruses in

(20)

6

causing mild illness. However, their overrepresentation in the asymptomatic or mildly symptomatic adult population cohorts, supports the notion that detection of such viruses in hospitalized patients should be interpreted with caution. This has also been emphasized by studies in children showing frequent detection of RV in asymptomatic children and long persistence of detectable viral RNA after acute infection [5] [11] [12] [13] [14] [18] [19].

Additional parameters in aid of clinical interpretation of diagnostic results are therefore desired. Viral load might represent one such parameter. In accordance with previous reports, we found significant correlations between viral load and illness severity, as approximated by symptom status and health care use, for InfA and RSV, indicating a relation between levels of viral replication and disease outcome [3] [5] [20] [21] [22] [23] [24] [25]. However, these correlations were less clear for viruses for which additional interpretative parameters are most needed. No significant correlations of RV and hCoV loads, as estimated by C_pvalues, were observed across severity groups. In contrast to previous observations in children [5], RV loads seemed higher in asymptomatic or mildly symptomatic adults in the general population when compared to hospitalized patients, with the exception of those admitted to the ICU in whom C_pvalues for RV were similar to C_p values in the general population. Interestingly, when only considering patients presenting at or admitted to the hospital, highest viral loads for both RV and hCoV were observed in ICU-admitted patients and a significant correlation between C_pvalues and approximated illness severity was observed for RV. Based on the latter observation, it is tempting to speculate that these viruses may well contribute to severe disease necessitating intensive care in certain situations, for instance, in the presence of comorbidities or other risk factors. To further our understanding, investigating the presence of these viruses in the lower respiratory tract of these patients is essential.

The strength of our study is that we were able to investigate the relative distribution and viral load of respiratory viruses in the general population and compare these to available routine diagnostic results from hospitalized patients in the same catchment area during the same period. A major limitation inherent to this approach is the absence of more detailed patient and clinical information from the hospital population, such as ethnicity, smoking history and the presence of comorbidities, which prevented us from correcting for these factors or identifying certain risk groups. In addition, the use of symptom status and health care use as a proximate for illness severity is rather

(21)

crude. Finally, we were able to perform meaningful analyses for the most frequently detected viruses only and larger studies are needed to gain insights into the role of less prevalent viruses. However, despite these limitations, our observations support the roles of influenza viruses, RSV and hMPV as important causes of adult morbidity necessitating hospital care. While detailed prospective studies across the different spectra of disease and infectious etiologies are clearly needed, our observations also provided direction towards understanding the contribution of less pathogenic viruses, such as RV, hCoV and hBoV, to severe acute respiratory infections.

Financial support

The HELIUS study is conducted by the Academic Medical Center Amsterdam and the Public Health Service of Amsterdam. Both organizations provided core support for the study. The HELIUS study is also funded by the Dutch Heart Foundation, the Netherlands Organization for Health Research and Development (ZonMw), and the European Union (FP-7). Parts of this work also received funding from the European Union Seventh Framework Programme (FP7) under the project PREPARE (grant agreement No 602525).

Acknowledgements

We are very grateful to the participants of the HELIUS study and the management team, research nurses, interviewers, research assistants and other staff who have taken part in collecting the data of the HELIUS study.

(22)

6 References

1- Woolpert T, Brodine S, Lemus H, Waalen J, Blair P, Faix D. Determination of clinical and demographic predictors of laboratory-confirmed influenza with subtype analysis. BMC Infect Dis 2012;12:129.

2- Jartti T, Jartti L, Petola V, Waris M, Ruuskanen O. Identificationof respiratory viruses in asymptomatic subjects: asymptomatic respiratory viral infections. Pediatr Infect Dis J 2008;27:1103-7.

3- DeVincenzo JP, Wilkinson T, Vaishnaw A, et al. Viral load drives diseases in humans experimentally infected with respiratory syncytial virus. Am J Respir Crit Care Med 2010; 182:1305–14.

4- Do LA, van Doorn HR, Bryant JE, et al. A sensitive real-time PCR for detection and subgrouping of human respiratory syncytial virus. J Virol Methods 2011;179:250–5.

5- Jansen RR, Wieringa J, Koekkoek SM, et al. Frequent detection of respiratory viruses without symptoms: toward defining clinically relevant cutoff values. J Clin Microbiol 2011;49:2631-6.

6- Jansen RR, Schinkel J, Dek I, et al. Quantitation of respiratory viruses in relation to clinical course in children with acute respiratory tract infections. Pediatr Infect Dis J 2010;29:82-4.

7- de Jong MD, Simmons CP, Thanh TT, et al. Fatal outcome of human influenza A (H5N1) is associated with high viral load and hypercytokinemia. Nat Med 2006;12:1203-7.

8- Stronks K, Snijder MB, Peters RJ, Prins M, Schene AH, Zwinderman AH. Unravelling the impact of ethnicity on health in Europe: the HELIUS study. BMC Public Health 2013;13:402.

9- Human Tissue and Medical Research: Code of conduct for responsible use (2011). Available at: http:// www.federa.org/sites/default/files/digital_version_first_part_code_of_conduct_in_uk_2011_12092012.pdf. Accessed 25 Februari 2015.

10- Graat JM, Schouten EG, Heijnen ML, et al. A prospective, community-based study on virologic assessment among elderly people with and without symptoms of acute respiratory infection. J Clin Epidemiol 2003;56:1218-23.

11- Makela MJ, Puhakka T, Ruuskanen O, et al. Viruses and bacteria in the etiology of the common cold. J Clin Microbiol 1998;36:539–42.

12- Winther B, Hayden FG, Hendley JO. Picornavirus infections in children diagnosed by RT-PCR during longitudinal surveillance with weekly sampling: Association with symptomatic illness and effect of season. J Med Virol 2006;78:644–50.

13- Van Benten I, Koopman L, Niesters B. Predominance of rhinovirus in the nose of symptomatic and asymptomatic infants. Pediatr Allergy Immunol 2003;14:363–70.

14- Nokso-Koivisto J, Kinnari TJ, Lindahl P, Hovi T, Pitkaranta A. Human picornavirus and coronavirus RNA in nasopharynx of children without concurrent respiratory symptoms. J Med Virol 2002;66:417–420. 15- Van den Hoogen BG, van Doornum GJ, Fockens JC, et al. Prevalence and clinical symptoms of human

metapneumovirus infection in hospitalized patients. J Infect Dis 2003;188:1571–77.

16- Mullins JA, Erdman DD, Weinberg GA, et al. Human metapneumovirus infection among children hospitalized with acute respiratory illness. Emerg Infect Dis 2004;10:700–5.

17- Robinson JL, Lee BE, Bastien N, Li Y. Seasonality and clinical features of human metapneumovirus infection in children in northern Alberta. J Med Virol 2005;76:98–105.

18- Singleton RJ, Bulkow LR, Miernyk K, et al. Viral respiratory infections in hospitalized and community control children in Alaska. J Med Virol 2010;82:1282–90.

19- Freymuth F, Vabret A, Cuvillon-Nimal D, et al. Comparison of multiplex PCR assays and conventional techniques for the diagnostic of respiratory virus infections in children admitted to hospital with an acute respiratory illness. J Med Virol 2006;78:1498–504.

20- Fuller J.A. Njenga MK, Bigogo G, et al. Association of the CT Values of Real-Time PCR of Viral Upper Respiratory Tract Infection With Clinical Severity, Kenya. J Med Virol 2013;85:924–32.

21- Lee N, Chan PKS, Hui DSC, et al. Viral loads and duration of viral shedding in adult patients hospitalized with influenza. J Infect Dis 2009;200:492–500.

22- Ahmed JA, Katz MA, Auko E, et al. Epidemiology of respiratory viral infections in two long-term refugee camps in Kenya, 2007-2010. BMC Infect Dis 2012;17:12-7.

(23)

23- Wurzel DF, Marchant JM, Clark JE, et al. Respiratory virus detection in nasopharyngeal aspirate versus bron-choalveolar lavage is dependent on virus type in children with chronic respiratory symptoms. J Clin Virol 2013;58:683-8.

24- Duncan CB, Walsh EE, Peterson DR, Lee FE, Falsey AR. Risk factors for respiratory failure associated with respiratory syncytial virus infection in adults. J Infect Dis 2009;200:1242–6.

25- Houben ML, Coenjaerts FE, Rossen JW, et al. Disease severity and viral load are correlated in infants with primary respiratory syncytial virus infection in the community. J Med Virol 2010; 82:1266–71.

(24)

6

Table S1. Associations betw een appr oximat ed illness se

verity and PCR positi

vity , b y type o f vir us Vir us and gr oups b y appr oximat ed illness se verity PCR positi ve/ total (%)* Cr ude OR P-value Adjust ed OR 1 P-value An y vir us Asympt omatic HELIUS 16/291 (6%) 1 <0.001 1 <0.001 Sympt omatic HELIUS 54/297 (18%) 3.8 (2.1-6.8) 3.7 (2.0-6.6) Outpatients 41/116 (35%) 9.4 (4.9-17.7) 9.4 (4.9-17.9) Inpatients 80/284 (28%) 6.7 (3.8-11.9) 6.7 (3.7-12.0) ICU patients 66/200 (33%) 8.5 (4.7-15.2) 9.1 (5.0-16.7) Test f or tr end (P -v alue) <0.001 Rhino vir us Asympt omatic HELIUS 6/291 (2%) 1 0.010 1 0.007 Sympt omatic HELIUS 22/297 (7%) 3.8 (1.5-9.5) 3.6 (1.4-9.2) Outpatients 13/116 (11%) 5.9 (2.2-16.2) 6.7 (2.4-18.5) Inpatients 23/284 (8%) 4.2 (1.7-10.4) 4.5 (1.8-11.6) ICU patients 13/200 (7%) 3.3 (1.2-8.8) 3.1 (1.1-8.7) Test f or tr end (P -v alue) 0.059 H uman c or ona vir us Asympt omatic HELIUS 5/291 (2%) 1 0.112 1 0.123 Sympt omatic HELIUS 15/297 (5%) 3.0 (1.1-8.5) 3.1 (1.1-8.7) Outpatients 8/116 (7%) 4.2 (1.4-13.2) 3.9 (1.2-12.8) Inpatients 12/284 (4%) 2.5 (0.9-7.3) 2.3 (0.8-7.0) ICU patients 12/200 (6%) 3.7 (1.3-10.5) 3.8 (1.3-11.2) Test f or tr end (P -v alue) 0.062

Supplement

(25)

Table S1 (c ontinue d). Associations betw een appr oximat ed illness se

verity and PCR positi

vity , b y type o f vir us Vir us and gr oups b y appr oximat ed illness se verity PCR positi ve/ total (%)* Cr ude OR P-value Adjust ed OR 1 P-value In fluenza A vir us 2 Asympt omatic HELIUS 0/291 (0)

}

1 <0.001 1 <0.001 Sympt omatic HELIUS 6/297 (2%) Outpatients 4/116 (3%) 3.5 (0.9-12.5) 3.8 (1.0-14.4) Inpatients 26/284 (9%) 9.8 (3.9-24.0) 10.7 (4.1-27.8) ICU patients 20/200 (10%) 10.8 (4.3-27.2) 13.4 (5.0-35.6) Test f or tr end (P -v alue) <0.001 Respir at or y syncytial vir us Asympt omatic HELIUS 2/291 (1%) 1 0.010 1 0.015 Sympt omatic HELIUS 4/297 (1%) 1.9 (0.4-10.9) 2.1 (0.4-11.9) Outpatients 8/116 (7%) 10.7 (2.2-51.2) 11.5 (2.3-57.3) Inpatients 11/284 (4%) 5.8 (1.3-26.5) 6.0 (1.3-28.7) ICU patients 6/200 (3%) 4.5 (0.9-22.4) 3.9 (0.8-20.5) Test f or tr end (P -v alue) 0.028 *U nless stat ed other wise AMC= A cademic M edic al C ent er HELIUS= H ealth y Li fe I n an U rban S etting ICU = I nt ensi ve c ar e unit OR = odds r atio PCR = pol ymer ase c hain r eaction 1 Sex, ag e, month and y ear o f sampling w er e a priori inc luded in multi variable anal ysis 2 Asympt

omatic and sympt

omatic HELIUS par

tici

pants w

er

e c

ombined in uni

variable and multi

variable anal ysis , bec ause in fluenza A vir us w as not det ect ed among asympt

omatic HELIUS par

tici

pants

(26)

6

Table S2. Linear r egr ession anal ysis o

f the association betw

een appr oximat ed illness se verity and cr ossing point v alues o f positi ve PCR r esults , b y type o f vir us

HELIUS & AMC gr

oups AMC gr oups Vir us and gr oups b y health c ar e exposur e Cr ude β Adjust ed β 1 Cr ude β Adjust ed β 1 Rhino vir us Asympt

omatic & sympt

omatic (n=28) 1 1 Outpatients (n=13) 2.7 (1.2 t o 4.2) 2.9 (1.2 t o 4.6) 1 1 Inpatients (n=23) 2.7 (1.5 t o 3.9) 2.6 (1.1 t o 4.2) -0.02 (-1.4 t o 1.4) -0.2 (-1.6 t o 1.2) ICU patients (n=13) -0.7 (-2.1 t o 0.8) -0.9 (-2.6 t o 0.8) -3.4 (-4.9 to -1.8) -3.7 (-5.4 t o -2.1) H uman c or ona vir us Asympt

omatic & sympt

omatic (n=20) 1 1 Outpatients (n=8) 1.6 (-2.2 t o 5.4) 2.1 (-2.7 t o 6.8) 1 1 Inpatients (n=12) 0.4 (-2.9 t o 3.8) 0.5 (-3.4 t o 4.4) -1.2 (-5.3 t o 2.9) -0.6 (-5.5 t o 4.4) ICU patients (n=12) -0.5 (-3.8 t o 2.8) 0.2 (-3.9 t o 4.3) -2.1 (-6.2 t o 1.9) -0.6 (-6.2 t o 5.0) In fluenza A vir us Asympt

omatic & sympt

omatic (n=6) 1 1 Outpatients (n=4) -1.1 (-5.6 t o 3.4) -2.9 (-7.8 t o 1.9) 1 1 Inpatients (n=26) -3.0 (-6.2 t o 0.2) -3.9 (-7.4 t o -0.5) -1.9 (-5.6 t o 1.8) -1.1 (-4.9 t o 2.8) ICU patients (n=20) -4.4 (-7.6 t o -1.1) -5.6 (-9.2 t o -2.0) -3.2 (-7.1 t o 0.7) -2.9 (-6.9 t o 1.0) Respir at or y syncytial vir us Asympt

omatic & sympt

omatic (n=6) 1 1 Outpatients (n=8) 1.6 (-3.1 t o 6.4) -1.0 (-8.2 t o 6.2) 1 1 Inpatients (n=11) -1.2 (-5.7 t o 3.3) -4.1 (-10.7 t o 2.5) -2.8 (-7.3 t o 1.7) -3.0 (-8.7 t o 2.6) ICU patients (n=6) -4.0 (-9.1 t o 1.1) -6.7 (-13.6 t o 0.3) -5.7 (-10.9 t o -0.5) -5.8 (-12.9 t o 1.4) 1 Sex, ag e, month and y ear o f sampling w er e a priori inc luded in multi variable anal ysis AMC= A cademic M edic al C ent er HELIUS= H ealth y Li fe I n an U rban S etting ICU = I nt ensi ve c ar e unit PCR = pol ymer ase c hain r eaction β = beta, repr

esenting the unstandar

diz

ed slope o

f the r

egr