Effectiveness of Immunization Products Against Medically Attended Respiratory Syncytial Virus Infection: Generic Protocol for a Test-Negative Case-Control Study Free

Abstract

Monitoring the real-life effectiveness of respiratory syncytial virus (RSV) products is of major public health importance. This generic protocol for a test-negative design study aims to address currently envisioned approaches for RSV prevention (monoclonal antibodies and vaccines) to study effectiveness of these products among target groups: children, older adults, and pregnant women. The generic protocol approach was chosen to allow for flexibility in adapting the protocol to a specific setting. This protocol includes severe acute respiratory infection (SARI) and acute respiratory infection (ARI), both due to RSV, as end points. These end points can be applied to studies in hospitals, primarily targeting patients with more severe disease, but also to studies in general practitioner clinics targeting ARI.

Human respiratory syncytial virus (RSV) is the most common pathogen identified in young children with lower respiratory tract infections (LRTI), a form of acute respiratory infection (ARI). It was estimated that globally, in 2019, there were 3.6 million RSV-associated acute LRTI hospital admissions and approximately 100 000 RSV-attributable deaths in children aged 0–60 months [1]. Most children are infected with RSV during the first 2 years of life and reinfections are very common. Clinical manifestations of RSV infection range from mild upper respiratory tract infections to severe LRTI. The most common form of RSV-associated LRTI among infants is bronchiolitis [2]. In infants, risk factors for hospitalization are preterm birth, bronchopulmonary dysplasia, hemodynamically significant congenital heart disease, low birth weight, and immunodeficiencies [2]. In older adults, RSV is also a common cause of ARI, with hospitalization and mortality rates comparable to influenza [3]. RSV infections can lead to exacerbation of existing chronic conditions, particularly chronic obstructive pulmonary disease and congestive heart failure [4, 5]. Hence, RSV is an important risk factor for hospitalization in adults with underlying comorbidities and frail people, such as people living in nursing homes [6, 7].

There is a single RSV serotype with 2 major antigenic subgroups, A and B. These 2 types may circulate simultaneously in the population, but generally 1 of the subtypes predominates [2, 8, 9]. Originally, RSV followed a distinct seasonality with onset in late fall and a peak between December and February in temperate regions of the northern hemisphere [2]. However, in many European countries, there were low levels of RSV activity during the 2020/2021 season [10, 11]. In some countries, this period of low activity was followed by a high peak of RSV activity during the summer [12]. This extraordinary pattern is likely due to the restrictions during the coronavirus disease 2019 (COVID-19) pandemic and will probably be restored, but warrants year-round surveillance.

Clinical symptom severity is highly correlated with viral load. RSV viral load significantly declines with increasing duration of symptoms before collection of specimens [13, 14]. In studies investigating the relationship between clinical score and RSV load in hospitalized children, it was found that 5 days after admission, clinical score and viral load significantly decreased [15, 16]. In another study, 5 days after admission, most respiratory symptoms either resolved or decreased in severity [17]. In adults, viral load peaks at 6 days postinoculation and steeply declines after 7 days [18]. However, polymerase chain reaction (PCR) tests may detect RSV up to 10 days after symptom onset [19, 20, 21].

Currently, the monoclonal antibody (mAb) palivizumab is approved for use in high-risk infants as prophylaxis [9]. In the European Union, the mAb nirsevimab is also approved for use in newborns and children since 31 October 2022 [22]. Since June 2023, the vaccine Arexvy (Respiratory Syncytial Virus Vaccine, Adjuvanted) is also authorized in the European Union, for use in older adults [23]. Many more immunization products are being tested and will likely become available for use in the coming years [24, 25].

To estimate immunization product effectiveness, a test-negative design (TND) study protocol was developed. This generic protocol is not intended to be implemented directly. Instead, the goal of this protocol is to assist researchers to draft site-specific protocols in the following years during the postmarketing phase. The term immunization is used because this protocol can be used both for the evaluation of monoclonal antibodies and vaccines.

OBJECTIVES

The primary objective is to estimate RSV immunization effectiveness (IE) against medically attended, laboratory-confirmed (severe) RSV infection in infants, children or older adults. The secondary objectives are to:

• estimate IE against medically attended, laboratory-confirmed subtype A and B RSV infection, and by genotype (obtained by whole-genome sequencing [WGS]) RSV infection;

• study factors affecting RSV IE, including age, sex, RSV season period (start or peak or end of the RSV season), comorbidity, gestational age at birth (in case of infants), maternal RSV vaccination status (in case of infants), gestational timing of maternal vaccination (in case of infants), time since onset of symptoms, and time since immunization;

• estimate IE against laboratory-confirmed RSV infection for specific brands or product types (in case multiple brands are used at the same time);

• estimate IE against hospital-acquired RSV infection (hospitalized patients testing positive for RSV >48 hours after hospital admission).

METHODS

Study Design

This concerns an observational test-negative case-control study with RSV testing for enrolled subjects. The study could take place in primary care or hospital setting. The study setting is defined by each study site depending on the research needs and available data. Subjects should be recruited from similar settings for a given study or analysis. Each study patient with ARI or hospitalized with severe acute respiratory infection (SARI) should be tested for the detection of, at least, RSV, influenza, and severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2).

If possible, the study should take place within an integrated, year-round, respiratory pathogen surveillance system, as recommended by Teirlinck et al [26]. A year-round surveillance is warranted as the seasonality may vary. The site-specific study period for analysis will start when the RSV season begins (based on pathogen surveillance, eg, when the number of detections is above a certain threshold [27]) in the country or region and will finish at the end of the RSV season (also based on pathogen surveillance). Conducting the study during the RSV season will allow for adequate sampling of both cases and controls.

Study Population

The study population consists of patients who:

1. are eligible to receive an mAb or an RSV vaccination and who seek care for a (severe) acute respiratory infection (SARI/ARI);

2. have no contraindication for the RSV immunization product being studied; and

3. have respiratory testing completed for RSV, influenza, and SARS-CoV-2.

The definitions below are based on the official World Health Organization case definitions for RSV surveillance but a few adaptations were made (the period for acute disease was changed from 10 to 7 days to be in line with the recommendations for testing, and the definitions of respiratory infection in ARI/SARI were aligned). As stated above, these case definitions can be adapted for site-specific study populations, although the use of these official, global definitions will facilitate comparison of studies.

ARI Definition

A patient with an ARI [28] is defined as a person with:

• acute disease—defined as sudden onset (within the last 7 days) or worsening of symptoms; and

• respiratory infection—defined as having at least 1 of the following:

  • ○shortness of breath

  • ○cough

  • ○sore throat

  • ○coryza.

SARI Definition

A patient with a SARI [28] is defined as a person with:

• severe disease—defined as requiring hospitalization; and

• acute disease—defined as onset within the last 7 days; and

• respiratory infection—defined as having at least 1 of the following:

  • ○shortness of breath

  • ○cough

  • ○sore throat

  • ○coryza.

In infants younger than 6 months, SARI additionally includes those who present with:

• apnea (defined as temporary cessation of breathing from any cause); and/or

• sepsis—defined as

  • ○fever (37.5°C or above) or hypothermia (less than 35.5°C), and

  • ○shock (lethargy, fast breathing, cold skin, prolonged capillary refill, fast weak pulse), and

  • ○seriously ill with no apparent cause.

Inclusion Criteria

Patients are eligible if they:

• fulfil age-specific criteria of the site-specific study population;

• present in any health care setting with respiratory symptoms;

• meet the ARI/SARI case definition;

• are eligible for immunization;

• accept to participate; and

• provide informed consent (or their parents/caregivers provide informed consent) according to country procedures.

Exclusion Criteria

Patients will not be enrolled in the study if they:

• have a contraindication for the RSV immunization due to a condition listed in the summary of product characteristics;

• had a respiratory specimen taken ≥7 days after ARI/SARI onset;

• were unable or unwilling to provide the protocol required respiratory specimen(s);

• were previously enrolled in the study within the same RSV season;

• were not immunized within the appropriate timeframe for the product to be effective (dependent on the product characteristics; usually 14 days for vaccines) before disease onset.

In case the objective is to determine effectiveness of immunization during pregnancy, the infants whose mothers have received RSV immunization before pregnancy or within 14 days before birth should be excluded.

Reasons for exclusion should be documented.

Recruitment and Eligibility

A systematic screening of all patients with respiratory symptoms admitted to the hospital or health care-seeking in the outpatient clinic should be organized. Patients admitted to the hospital or presenting at the general practitioner (GP) site who meet the case definition for ARI/SARI and the inclusion criteria are proposed to be part of the study and provide informed consent, depending on the regulations at the site. The ICD-10 (hospital) or ICPC-2 (GP) codes listed in Supplementary Table 1 could be used as a guideline to identify potential study participants, if appropriate at specific sites.

Outcome

The outcome of interest is medically attended laboratory-confirmed (severe) RSV infection.

Cases are defined as persons with a laboratory-confirmed (severe) RSV-positive infection—a person in the study population, meeting the (S)ARI case definition with at least 1 respiratory sample positive for RSV. Controls are persons with a laboratory-confirmed RSV-negative infection—a person in the study population, meeting the (S)ARI case definition with all respiratory samples being negative for RSV.

Matching of cases and controls could be considered to improve statistical efficiency and ensure similarity between cases and controls on key confounding variables. In case of matching, the following factors could be used: age, prematurity (for infants), frailty (for older adults), comorbidities, and calendar time.

Laboratory Confirmation

Study staff will collect respiratory specimens from all eligible patients. Respiratory sampling should preferably be done from both the nasopharynx and oropharynx, using a stiff oropharynx swab or a flexible nasopharynx swab [29]. Among older adults, collection of more than 1 sample type should be considered if resources allow, because this would reduce the number of undetected infections in the control group [30, 31]. In children, adding multiple specimens (eg, sputum or oropharyngeal swab in addition to nasopharyngeal swab) results in a (nonsignificant) increase in RSV detection [32].

The diagnosis of RSV infections is based on nucleic acid amplification tests (NAAT; eg, quantitative real-time reverse transcriptase-polymerase chain reaction [qRT-PCR], with 100% sensitivity and specificity [33, 34]) or molecular point-of-care (mPOC) tests. The current generation of mPOC tests has a high sensitivity (97%–100%) and specificity (99%–100%) across all ages and yields results very quickly [28, 35–37].

RSV laboratory confirmation should ideally be done using NAATs able to identify RSV subtype A and B, preferably in a laboratory setting, because of the possibility of further analyzing the respiratory specimen at a later stage and, thus, allowing for detection of other pathogens and WGS. If mPOC tests are used, they should at least be able to identify influenza and SARS-CoV-2 as well [26]. Ideally, a full viral respiratory panel would be conducted if resources allow. Inclusion of infection with other vaccine-preventable pathogens on the control population has been shown to bias influenza and SARS-CoV-2 vaccine effectiveness (VE) estimates downwards due to correlated vaccine uptake, and extended pathogen testing would allow for use of only nonvaccine-preventable pathogens as controls, which is one recommended pathway to reduce this bias [38]. The selection procedures for WGS of RSV samples as well as the laboratory procedures (type and number of samples taken, storage, transport) and tests used should be clearly described.

Exposure

The exposure of interest is immunization with any RSV immunization product, as specified in Table 1. RSV immunization can be ascertained based on:

• registration of immunization in the electronic patient dossier;

• registration of immunization in an immunization registry;

• evidence of immunization through pharmacy delivery or reimbursement;

• registration of immunization on/in vaccination card/vaccination booklet;

• self-report of immunization.

Objective measures of immunization are considered highly preferable to self-reported measures and should be the primary approach to immunization exposure assessment.

Covariables

Information to be collected includes the covariates listed in Supplementary Table 2, where items with an asterisk are crucial and other items are preferable. For each study population, a different set of the covariates presented in Supplementary Table 2 will be relevant. The methods used to collect these covariates should be described. All covariates are defined prior to the beginning of study period except for vaccination against influenza, pneumococci, or COVID-19, which are time-dependent variables.

Sample Size Calculation

The number of RSV infection cases needed in a single TND study to detect a crude IE different from 0 was calculated for a range of study parameters [39]. The calculations were performed for each combination of the following study parameters and the required number of RSV cases can be found in Supplementary Table 3.

1. IE of 50% or 80%;

2. control to case ratio of 1, 2, or 4 (the case to control ratio depends on the prevalence and virulence of the disease and on the participation rate of the source population with or without the disease of interest);

3. overall immunization coverage among control subjects of 30%, 50%, or 60%.

Data Analysis

For the primary analysis, symptoms should not have started ≥48 hours after a hospital admission to prevent the inclusion of nosocomial infections. Patients with nosocomial infections should be analyzed separately from patients with community-acquired infections.

A separate statistical analysis plan for each site-specific study should be made; it must be finalized prior to any analyses and describe analyses and a priori decision rules. Variables that can be evaluated for missingness should be specified in this plan and imputation rules must be predefined. The proportion of eligible ARI/SARI patients who are accepted to participate in the study will be calculated (response rate).

Demographics and Baseline Characteristics

The baseline characteristics of the study participants are presented and can be described and tabulated for cases and controls separately and for immunized and nonimmunized subjects within each group (by brand and overall).

Measure of Effect

IE is computed as IE = (1 − OR) × 100%, where OR denotes the odds ratio, comparing the odds of immunization among the cases to the odds of immunization among controls. A 95% confidence interval is computed around the point estimate.

The analyses should take into account potential confounding factors and effect modifiers. Analysis may be stratified according to (if sample size allows):

• age groups (within infants or elderly could be stratified);

• time: calendar time will be used as a proxy for the circulating RSV variant; using systematic samples of the sequenced isolates, the proportion of virus changes will be calculated at different time points along the season (eg, by week or group of weeks in the season);

• presence of at least 1 underlying condition;

• time of respiratory sampling: < 5 days or ≥5 days after symptom onset;

• subtype of RSV (A or B);

Where sample size allows, further analyses may be carried out. These include:

• IE at different time points along the season (eg, IE by week or group of weeks in the season [IE for weeks 2–3, 4–5, 6–7, etc.]);

• IE by time since immunization: time since immunization can be calculated by subtracting the date of immunization from the date of onset and then be modelled as a continuous or categorical variable;

• a sensitivity analysis in patients who tested positive for RSV in the current epidemic season before the onset of symptoms leading to the current episode of health care-seeking (eg, patients who have their second or third laboratory-confirmed RSV infection);

• sensitivity analyses to assess potential bias of influenza or SARS-CoV-2 vaccination and/or testing (also) positive for influenza or SARS-CoV-2, eg, a sensitivity analysis where only those who test positive for other respiratory pathogens are controls; for older adults, this analysis would also hopefully reduce inclusion of RSV-infected persons in the control population who tested negative due to collection of a single or limited sample types, which could bias VE estimate towards the null;

• assessment of the different genotypes to gain insight into the degree of variation among circulating strains and be informative about changes in the characteristics of the virus;

• sensitivity analysis in patients with self-reported immunization could be considered;

• sensitivity analysis in patients who received more than 1 type of RSV prophylaxis prior to/during the season (eg, maternal immunization too close to birth plus nirsevimab after birth, or were involved in an RSV mAb or RSV vaccine trial, or received RSV immunization in a previous season).

Missing Data

Any missing data should be documented. If many data are missing and/or there is evidence of bias in the missing data, multiple imputation (at study level) can be done.

DISCUSSION

This generic protocol for a test-negative case-control study is developed to study RSV immunization product effectiveness. It includes both SARI and ARI due to RSV as end points, allowing for studies at hospitals, targeting patients with more severe disease, but also for studies at GP clinics. Depending on the site and setting, specific outcomes within the (S)ARI definition can be defined, such as LRTI. For each outcome, for example, hospitalization or mild infection, a separate study should be performed as only 1 outcome can be used in TND studies, dependent on the case definition.

A TND study is a widely used and well-established method for assessing VE against influenza and SARS-CoV-2 [40, 41]. TND studies provide valid estimates of VE against influenza and have smaller bias than estimates from traditional case-control studies. VE estimates of TND influenza studies are more reliable when the outcome is medically attended influenza than symptomatic influenza if vaccination reduces symptom severity and, hence, vaccinated people are less likely to seek medical care than unvaccinated people [42, 43]. This reasonably applies to TND for RSV as well.

Strengths of the TND are that it reduces bias by enrolling cases and controls with the same health care-seeking behavior, it improves study efficiency, it can often be conducted within existing surveillance systems, and although it requires a highly specific and sensitive test, it does not (necessarily) depend on standard clinical testing as testing can also be performed only in the context of the study. Limitations of this design are that IE can be biased due to imperfect sensitivity and specificity of rapid diagnostic tests and that collection of valid immunization data can be challenging. Furthermore, as in any observational study, it is not possible to rule out the presence of unmeasured confounding [40]. Also, health care-seeking behavior may differ by country depending on the case management strategy (eg, recommendation of not going to the GP). The representativeness of cases included in the study may affect the IE estimates.

This protocol is intended to provide guidance for national and regional health authorities in the evaluation of implementation of RSV immunization products. Based on the surveillance or health care characteristic, a TND study approach or a cohort study approach can be chosen to assess RSV IE, whichever is most applicable. The generic protocol of the cohort study by Poukka et al [44] complements this TND protocol.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online (http://jid.oxfordjournals.org/). Supplementary materials consist of data provided by the author that are published to benefit the reader. The posted materials are not copyedited. The contents of all supplementary data are the sole responsibility of the authors. Questions or messages regarding errors should be addressed to the author.

Notes

PROMISE investigators. Harish Nair and Harry Campbell (University of Edinburgh, Edinburgh, UK); Louis Bont (University Medical Centre Utrecht, the Netherlands); Adam Meijer, Anne C. Teirlinck, Mirjam Knol, David Gideonse, Anoek Backx, Hester de Melker, Caren van Roekel, and Lance Presser (National Institute for Public Health and the Environment, Bilthoven, the Netherlands); Topi Turunen, Hanna Nohynek, Eero Poukka, and Annika Saukkoriipi (Finnish Institute for Health and Welfare, Finland); John Paget, Jojanneke van Summeren, and Michel Dückers (Netherlands Institute for Health Services Research, Utrecht, Netherlands); Terho Heikkinen (Department of Pediatrics, University of Turku and Turku University Hospital, Turku, Finland); Berta Gumí Audenis and Maica Llavero (Teamit Research SL, Barcelona, Spain); Leyla Kragten and Lies Kriek (Respiratory Syncytial Virus Network ReSViNET, the Netherlands); Kristýna Faksová, Michele Giardini, Hanne-Dorthe Emborg, and Francesca Rocchi (Statens Serum Institut, Copenhagen, Denmark); Cintia Muñoz Quiles, Javier Diez-Domingo, Charlotte Vernhes, Clarisse Demont, Aurelie Robin, David Neveu, Lydie Marcelon, Mathieu Bangert, Rolf Kramer, Oliver Martyn, Corinne Bardone, Vanessa Remy, and Sandra Chaves (Sanofi Pasteur SA, Lyon, France); Daniel Molnar, Gael dos Santos, Jean-Yves Pirçon, Bishoy Rizkalla, Elisa Turriani, Se Li, Noemie Napsugar Melegh, Philip Joosten, and Victor Preckler Moreno (GlaxoSmithKline Biologicals SA, London, UK); Aigul Shambulova, Arnaud Cheret, Delphine Quelard, Jeroen Aerssens, Karin Weber, Corinne Willame, Anna Puggina, Katherine Theis-Nyland, and Natalia Nikolayeva (Johnson & Johnson Medical, New Jersey, US); Veena Kumar, Hadi Beyhaghi, and Vivek Shinde (Novavax, Gaithersburg, MD, US); Beate Schmoele-Thoma, Elizabeth Begier, Kena Swanson, Tin Tin Htar, Jessica Atwell, Negar Aliabadi, Jen Deese, Deshayne Fell, Maria Maddalena Lino, and Monica-Flavia Turiga (Pfizer Ltd, New York City, NY, US); and Bahar Ahani (AstraZeneca, Cambridge, UK).

Financial support. This work was supported by the Innovative Medicines Initiative 2 Joint Undertaking under grant agreement number 101034339. This Joint Undertaking receives support from the European Union Horizon 2020 Research and Innovation Programme and the European Federation of Pharmaceutical Industries and Associations.

Supplement sponsorship. This article appears as part of the supplement “Preparing Europe for Introduction of Immunization Against RSV: Bridging the Evidence and Policy Gap.”

References

Author notes