Understanding group A streptococcal pharyngitis and skin infections as causes of rheumatic fever: protocol for a prospective disease incidence study

Group A Streptococcus (GAS) causes a broad spectrum of diseases, from superficial infections such as pharyngitis and impetigo, to invasive infections like necrotizing fasciitis and streptococcal toxic shock syndrome. GAS infections can also lead to the development of autoimmune diseases, such as acute rheumatic fever (ARF) [1, 2]. In an estimated 60% of ARF cases, carditis progresses to chronic rheumatic heart disease (RHD), which can produce permanent heart valve damage [3]. Unless further episodes of ARF are prevented with intramuscular injections of benzathine penicillin G every 28 days, ARF patients are likely to suffer worsening cardiac damage and increasing chances of heart failure, stroke and early death [4]. Globally approximately 34 million people are affected by RHD with an additional 47 million having asymptomatic damage to their heart values [5, 6]. Annually, severe GAS infections are responsible for an estimated 517, 000 deaths [3].

ARF was once common throughout the Western world, with entire hospital wards dedicated to caring for patients. Incidence rates declined sharply over the twentieth century, coinciding with improvements in living conditions and later the widespread use of antibiotics to treat streptococcal infections [7, 8]. Despite a world-wide decline, ARF and RHD remain important causes of morbidity and preventable early death in the low- and middle-income countries [3, 9]. ARF has also persisted in indigenous communities in some developed countries. In New Zealand, there are large and widening ethnic disparities in ARF rates [10, 11]. Ethic inequalities are particularly pronounced amongst Indigenous Māori and Pacific population groups living in New Zealand, with more than nine out of every 10 cases occurring in these groups [12]. The rate of new cases of ARF notified in New Zealand from July 2017 to June 2018 for Māori children aged 5–12-years was 26 per 100,000 and for Pacific children aged 5–12-years was 94 per 100,000, and for European/Other children was < 1 per 100,000 [12].

ARF largely occurs in children, particularly those living in socioeconomic deprivation. From July 2015 to June 2018, 64% of all new ARF cases in New Zealand were children aged 5–14-years [11], and between 2010 and 2013 people living in the most deprived areas of New Zealand were found to be 33 times more likely to develop ARF compared with those in the least deprived areas. Children living in the most deprived socioeconomic quintile have a 1:150 chance of being hospitalised with ARF by the time they reach the age of 15 years [9].

In response to the high rates of ARF and increasing ethnic inequalities, the New Zealand Ministry of Health introduced the Rheumatic Fever Prevention Programme (RFPP) in 2011. The RFPP placed a strong emphasis on sore throat management – that is, prompt detection by culturing throat swabs and if the swab was positive for GAS, a course of broad-spectrum antibiotics was provided to treat the infection promptly with the aim of preventing subsequent ARF [13, 14].

Treating pharyngitis caused by GAS infection with antibiotics is an important measure for preventing ARF. A major challenge to this approach is that it is extremely difficult to distinguish patients with true GAS pharyngitis from carriers who have viral pharyngitis when throat swab culturing alone is used to diagnose GAS pharyngitis [15, 16]. To date most studies investigating the incidence of GAS pharyngitis have identified cases using throat swabbing and culture. In some cases, a rapid antigen detection test (RADT) is used; however, laboratory culture is still considered the most specific means of GAS confirmation [17, 18]. Clinical symptoms are relatively non-specific so are often used in combination with RADT or culturing when diagnosing GAS pharyngitis [19‐21]. At present, the ‘gold standard’ method for diagnosis of true GAS pharyngitis utilises serological testing. This method measures the titres of antistreptolysin (ASO) and anti-DNase B (ADB) antibodies in two blood samples taken two-to-four-weeks apart [22] with seroconversion thought to distinguish true GAS pharyngitis from pharyngeal GAS carriers.

Only a small number of studies have investigated the incidence of serologically confirmed GAS pharyngitis. A recent meta-analysis assessed the prevalence of GAS pharyngitis and carriage in different settings [16]. The analysis identified 285 eligible studies and reported that in clinical settings; approximately 10% of children swabbed with a sore throat had serologically-confirmed GAS pharyngitis, with this increasing to around 50–60% when the child was GAS culture-positive. The analysis reported a prevalence of 10.3% (6.6–15.7%) for serologically-confirmed GAS pharyngitis in children from high-income countries and an asymptomatic GAS carriage prevalence of 10.5% (8.4–12.9%). Lower carriage prevalence was detected in children from low/middle income countries (5.9%, 4.3–8.1%) [16].

As most sore throats are caused by viral infection, it is likely that many patients considered to have pharyngitis caused by GAS are in fact GAS carriers with concurrent viral pharyngitis [23‐25]. Such individuals often receive false positive diagnoses and unnecessary antibiotic treatment, with carriers unlikely to benefit from antibiotic treatment. The Infectious Diseases Society of America makes a strong recommendation against routine antibiotic treatment of carriers [26]. Identifying the incidence of true (serologically confirmed) GAS pharyngitis is of critical importance when seeking to guide treatment, target interventions and inform vaccine development [23, 27].

There is growing evidence regarding the potential importance of GAS skin infections as a cause of ARF, either alone or in combination with GAS pharyngitis [1, 2, 28‐30]. In New Zealand rates of skin infections are markedly higher in New Zealand Māori and Pacific peoples compared with the New Zealand European/Other population groups [29]. Recent figures show that Māori and Pacific children under 9 years of age suffer disproportionately high rates of GAS skin infections in comparison to other ethnicities [31]. One New Zealand study conducted in the Tairawhiti region estimated an annual incidence rate of 106.7 cases per 1000 children; i.e. approximately one in nine children in the study region consulting their doctor for a skin infection each year [32]. Among skin infection cases hospitalised in that region, GAS was the second most isolated organism with 31% of children who had a microbiological swab taken positive for GAS; suggesting that this organism is a major contributor to such infections at a population level [33].

Identifying potentially modifiable risk factors for relevant GAS infections (pharyngitis and/or skin infections) is critical for the design of evidence-based ARF programmes. There is little published research in this area. One Australian study that has investigated risk factors for serologically confirmed GAS pharyngitis did not detect any association with crowding, gender, socio-economic status, or time spent at preschool, childcare or school [23]. Participating families were mainly small to moderate-sized and middle-class, so this study was not well powered to measure the effects of socio-economic deprivation [23], and therefore findings cannot be extrapolated to the groups at most risk of developing ARF in New Zealand.

The literature base of studies on risk factors for ARF and RHD is much larger than it is for GAS infections, though the studies are generally considered to be of poor quality [34]. One systematic literature review concluded that the findings supported a causal association between household crowding and low socioeconomic position and the development of ARF [34]. In New Zealand an ecological study found that the risk of ARF was associated with neighbourhood deprivation, household crowding, and the proportion of 5–14-year-old children living in the area [35].

The purpose of this research work is to fill critical gaps in scientific knowledge that are needed to support New Zealand’s Rheumatic Fever Prevention Programme (RFPP). This research will also add to our understanding of the pathophysiology of ARF. New Zealand invested more than $65 million in the RFPP, with most of these resources used to support large-scale sore throat management interventions. Unfortunately, we do not know what proportion of children treated through the RFPP actually had true GAS pharyngitis and much of the associated antibiotic use may have been unnecessary.

Similarly, we know little about the role of GAS skin infections, which may be an important precipitating event for ARF and thus might provide a key target for interventions. This study will also investigate the risk factors associated with GAS pharyngitis, carriage and skin infections, thus providing knowledge about potential prevention measures. The specific study aims are to:

The study design is a prospective disease incidence study that has an associated case-control study. The conditions of interest are divided into the following groups:

The study aims to include 1000 school aged (5–14-year-old) children in Auckland, 800 of whom have visited their healthcare professional, with the result of a throat or skin swab being sent to Labtests laboratory (sole diagnostic community laboratory service provider for the Auckland region) for investigation of presumed GAS infection. Families of children with a GAS positive throat or skin swab, or a GAS negative throat swab, will be phoned by Labtest’s microbiology staff and invited to take part in the study. In addition, 200 asymptomatic children who have previously taken part in the New Zealand Health Survey and consented to being contacted for further research will be recruited. CBG Health Research Limited, who provides independent public sector research services in New Zealand, will contact all Labtest referrals and potential asymptomatic participants (identified from previous participants in the New Zealand Health Survey) to arrange an initial home visit and specimen collection.

Figure 1 outlines the study recruitment, data gathering, specimen collection and laboratory analyses. Participants with the first four conditions of interest (n = 800) will have acute and convalescent serological testing for ASO/ADB titres (taken < 9 days post swab collection and 2–4 weeks post-acute blood specimen collection, respectively). Asymptomatic participants in the fifth group (n = 200) will have one blood specimen taken for ASO/ADB titre assays and one throat swab for microbiological culturing. Identification of GAS negative sore throat cases will be determined by routine culturing. Unequivocal GAS pharyngitis cases will be those participants who present with a GAS positive throat culture and a > 0.2 log-10 rise in ASO or ADB antibody titres between their acute and convalescent serum samples as previously described [23]. Probable GAS pharyngitis cases will include any participant where the first serum sample is taken after the acute phase, with ASO and/or ADB titres that are above the ULN determined for this study. Carriers will be determined by routine swab culturing that yields GAS but no increase in ASO or ADB titres, nor titres above the ULN. Skin infection cases will be determined using routine culturing of skin swabs that yield GAS.

A CBG researcher will visit the caregiver/parent and child in their home to explain the study. The researcher will provide the parent/caregiver and the child with a participant information sheet and if needed the researcher will read/explain this sheet. Information and consent sheets will be translated from English into Te Reo Māori, Cook Island Māori, Samoan, Tongan and Niuean. If language difficulties are encountered during the consent process, language translation services will be used.

Those participants who had an original positive GAS throat swab (n = 400) will receive a second home visitation, 2–4 weeks after the first home visit, where the CBG researcher will take a convalescent blood sample and a throat swab. Participants whose initial throat swab was GAS negative, and participants who had a positive GAS skin swab, do not require a throat swab to be collected and so will visit a Labtests collection centre, 2–4 weeks after their home visit, to have their second convalescent blood sample taken. An additional sample of n = 40 GAS throat swab positive cases with seroconversion will receive a third home visit from CBG staff, 6 months after study enrolment where a blood sample and a throat swab will be collected to investigate the persistence GAS antibody responses.

All specimens will be stored for a period of 10 years. The exact location of specimens (including details such as the freezer used for storage), will be recorded on an online track-and-trace database and eventually at a central location at the University of Otago, Wellington.

The study has both a Māori and Pacific Steering Group each comprised of a cross-section of key stakeholders, including researchers, public health workers, clinicians and community representatives (members are listed in the acknowledgements section). These groups reviewed the study questionnaire and protocol to ensure it was safe and appropriate. They also provide advice on cultural specific issues as arise during the study, and will be involved in interpretation and dissemination of research findings. One of the investigators (MH) is an active general practitioner in Auckland. The Study Manager is in ongoing communication with Primary Health Organisations in Auckland. These relationships ensure that the study can receive and incorporate feedback about the study from General Practitioners, nurses and other health care providers.

The Ministry of Health’s Northern-A Health and Disability Ethics Committee (HDEC) approved this study (reference /NTA/262). The Ministry of Health approved the study using participants from the New Zealand Health Survey who had consented to taking part in extra research.

A systematic review of GAS pharyngitis and pharyngeal carriage reported a pooled prevalence of serologically confirmed GAS infection amongst those with GAS positive throat swabs, of around 50%, for children recruited passively (i.e. where patients present to primary healthcare with a sore throat) [16]. The included GAS pharyngitis studies were based on demonstrating an ASO or ADB antibody titre rise of at least 0.2 log between acute and convalescent serum samples. Based on data from Gisborne, we could expect skin infection consultations at a doctors’ practice to be 107/1000/year [32] and GAS as the causative organisation in 31% [33] implying up to approximately 270 GAS skin infections a year within this population group.

For exposures that are relatively common, such as household crowding (58% in this population) [44], this study size will be able to detect an odds ratios (OR) for disease of 1.8 in exposed subjects relative to unexposed subjects with probability (power) of 0.8 and type I error probability of 0.05 using an uncorrected chi-squared statistic to evaluate the null hypothesis. Table 1 indicates the minimum protective and harmful effects we will be able to distinguish with 80% power when a factor of interest has certain prevalence difference in the control population compared with cases.

The goal of the data analysis is to support effective investigation of the aims and research questions. We will calculate the throat swabbing rate and GAS positive rate for our study population using Labtests laboratory data as the numerator and census population data as the denominator (as we have previously done for several regions of NZ) [45]. Since Labtests conducts all the testing for this population using standardised and consistent methods, this calculation should provide robust population rates. We will then apply the results of serological testing generated by this study (i.e. the proportion of GAS positive individuals that were carriers and those with pharyngitis), to estimate the prevalence/incidence of these conditions at a population level. We will then repeat these analyses for age strata and ethnic groups.

Our results will also allow us to estimate the test performance of GAS culture for detecting GAS pharyngitis in this population and context. We will use GAS serology as the gold standard to assess sensitivity, specificity, positive and negative predictive values of a GAS positive throat culture for detecting GAS pharyngitis.

Initial univariate analysis will compare the odds of exposure to each risk factor investigated for GAS pharyngitis and GAS skin infection cases compared with our various sets of control/comparison groups, with suitable statistical testing (for example Chi-square test for categorical variables). Multiple logistic regression will be used to calculate adjusted odds ratios. Non-additive interaction theoretically plausible or empirically supported interactions(s) between risk factors will be evaluated. At the end of these analyses, this study will be able to identify, with a reasonable degree of certainty, whether the risk of GAS pharyngitis and GAS skin infection in New Zealand is associated with a number of important risk factors that are potentially modifiable: structural and functional household crowding and bed sharing; poor indoor environments (cold, damp, mould) and fuel poverty; tobacco smoke exposure; limited resources for personal hygiene (e.g. washing, teeth cleaning); poor access to health services, including pharyngitis and skin infection treatment; and poor oral health.

This analysis will assess whether oral antibiotic treatment is as effective at clearing GAS carriage as well as GAS pharyngitis. This comparison will be based on re-swabbing all subjects who are GAS throat-swab positive at the second home visit. This is a clinically important and unanswered question in the field of GAS pharyngitis treatment.

Using ASO/ADB titres from asymptomatic children we will determine age-specific ULN values for children in New Zealand as well as carriage rates among asymptomatic children.

Multiple methods will be used to disseminate study findings. Findings will be communicated through the peer-reviewed literature and conference presentations with associated media coverage helping inform the wider public. At the conclusion of our study, our results will be made available as both printed and online resources.

We will disseminate our research findings in a clear, concise, and culturally-safe manner with Māori and Pacific communities. For example, we will develop press releases and provide these to Māori and Pacific media. We will participate in meetings and workshops which are relevant to the current study. This process will draw on the expertise of the Māori and Pacific Steering Groups. Members of the research group are already providing advice to the Ministry of Health and Health Promotion Agency of aspects of the RFPP programme and are recognised opinion leaders in clinical aspects of ARF and RHD management.