Introduction

Methods

This study is reported as per Consolidated Standards of Reporting Trials (CONSORT) guideline (S1 CONSORT Checklist) [10]. Details of PREVIX_COMBO and PREVIX_BOOST (including PREVIX_VOICES) protocols have been published [5,7]. Brief methods relevant to hearing outcomes from age 12 to 36 months are described below.

Results

The PREVIX_BOOST trial commenced participant recruitment in March 2013. Hearing tests were performed by the NT Government Hearing Services outreach program if possible. Research audiologists commenced in March 2017. Data collection was completed in September 2018.

At baseline age 12 months, 131 children were allocated to +P and 130 to +S; 1 child allocated to +P received +S and was included in the +P group for analysis. Of 461 hearing tests, 409 (89%) were VROA and 52 (11%) Play Audiometry; 74% were performed by research audiologists. Very few children were lost to follow-up for outcome measures other than hearing assessments.

There were no significant vaccine intergroup differences in characteristics of children in the hearing cohort at any age, other than at age 24 months when the +S group had a higher baseline proportion of siblings with “runny ears” (32%) than the +P group (13%), and at age 36 months when the +S group had a higher baseline proportion of breast feeding in infancy (73%) than the +P group (52%). (Table 1)

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Baseline characteristics and risk factors for the hearing cohort at each study visit age.

https://doi.org/10.1371/journal.pmed.1004375.t001

We report the numbers of children assessed by audiologists as a proportion of all children followed up within window by research nurses for any study outcome measure; only 23% PREVIX_BOOST children were age-eligible for hearing assessments at age 12 months, 35% at age 18 months, and around half (55%) at age 36 months. From 2017, for the additional study visits introduced at ages 24 and 30 months, 88% age-eligible children had hearing assessments at age 24 months, and 100% at 30 months (Table 2 and Fig 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Participant flow.

P: PCV13 13-valent pneumococcal conjugate vaccine. S: PHiD-CV10 10-valent pneumococcal H. influenzae protein D conjugate vaccine. *One child allocated to +P received +S.

https://doi.org/10.1371/journal.pmed.1004375.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Proportion of children with hearing loss by category (normal, mild, moderate, severe, >23·3 dB), and mean hearing loss (dB, SD) at ages 12 (baseline), 18, 24, 30, and 36 months, by booster dose (difference, 95% CI, p-value).

https://doi.org/10.1371/journal.pmed.1004375.t002

At baseline age 12 months, 32/131 (24%) children in the +P and 28/130 (22%) in the +S group had hearing tests; hearing loss was similar in both groups. (Table 2 and Fig 2) At age 18 months, 6 months post booster dose, 42 children in the +P group and 49 in the +S group had hearing tests; 36% (+P) and 16% (+S) had no hearing loss (difference 19%; (95% CI [2, 37], p = 0·05); 43% in each group had mild hearing loss, and 21% (+P) and 41% (+S) had moderate hearing loss (difference −19%; (95% CI [−38, −1], p = 0·07) (Table 2 and Fig 2). Mean hearing level was 26·2 dB (+P) and 29·6 dB (+S), (difference −3·4 dB; (95% CI [−6·7, −0·2] p = 0·04) (Table 2). The Hearing Australia threshold for recommended hearing assistance (mean hearing level (3fa) greater than 23·3 dB) [12] was met by 62% (+P) and 76% (+S) children (difference −14%; (95% CI [−33, 5] p = 0·18) (Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Hearing loss prevalence (none, mild, moderate, severe) by booster dose, at ages 12 (baseline), 18 (primary endpoint), 24, 30, and 36 months.

+P: booster dose of PCV13 13-valent pneumococcal conjugate vaccine. +S: booster dose of PHiD-CV10 10-valent pneumococcal H. influenzae protein D conjugate vaccine.

https://doi.org/10.1371/journal.pmed.1004375.g002

At ages 24, 30, and 36 months, the proportions of children with no hearing loss was higher, mean hearing level and moderate hearing loss tended to be lower in the +P group (Table 2 and Fig 2).

Analysis stratified by allocation to 3 primary course schedules in PREVIX_COMBO indicated that the lower prevalence of moderate hearing loss in the +P group was present in all subgroups (differences −8% between _PPP+P and _PPP+S, −5% between _SSS+P and _SSS+S, and −15% between SSSP+P and SSSP+S). The greatest difference was between SSSP+P (23%) and SSSP+S (38%) (difference −15%; (95% CI [−30, −0.1], p = 0·07) (Table 3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Prevalence of hearing loss in the +P versus +S groups, stratified by allocation to 3 primary course PCV schedules (_PPP, _SSS, and SSSP).

https://doi.org/10.1371/journal.pmed.1004375.t003

Fluctuating hearing loss was examined using data from children who had hearing tests at 2 or more consecutive 6-monthly assessments. Overall, at initial assessment, the proportion of hearing tests that found no hearing loss (35%), mild (36%), or moderate (29%) hearing loss was similar to 6-monthly follow-up (38%, 36%, 25%, respectively). However, transitions in hearing status were more dynamic than revealed by point prevalence: Of 73 children with no hearing loss initially, 51% had no hearing loss at follow-up, 37% transitioned to mild, and 12% to moderate hearing loss. Of 76 children with mild hearing loss initially, 31% transitioned to no hearing loss, 42% had persisting mild hearing loss, 26% transitioned to moderate hearing loss, and 1% to severe hearing loss. Of 61 children with moderate hearing loss initially, 33% transitioned to no hearing loss, 28% transitioned to mild hearing loss, and 39% had persisting moderate hearing loss (Table 4).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Transitions between consecutive 6-monthly hearing assessments.

https://doi.org/10.1371/journal.pmed.1004375.t004

Discussion

To our knowledge, this is the first study globally to evaluate hearing loss outcomes in an RCT of head-to-head or mixed schedule PCVs. This is also the first prospective longitudinal community-based study of scheduled hearing tests in First Nations children from age 12 to 36 months. A literature search for studies of either PCV13 or PHiD-CV10 that report either OM or hearing-related outcomes [14] was repeated for the period 2019 to 2023; of 54 identified studies, none reported hearing outcomes. A recent 2020 Cochrane review of PCVs for prevention of acute OM did not report hearing outcomes [16]. Other head-to-head RCTs compare immunogenicity of different PCV schedules or formulations [17,18] and impact on pneumococcal [19] and NTHi carriage [17], but impacts on hearing have not been included.

Unexpectedly, this report focussing on hearing outcomes found that, contrary to our hypotheses, the children assessed in the +P group had better hearing compared to the +S group at age 18 months. The −3 dB difference in mean hearing level is of unknown clinical value. However, the −19% difference in prevalence of moderate hearing loss and gain of 19% having no hearing loss is likely to benefit many First Nations children, particularly if this is sustained to age 36 months. In addition, around two-thirds of the children assessed at age 18 months met Hearing Australia criteria for hearing assistance [12], and the proportion was again lower in the +P group at age 18 months (62% versus 76%). At later ages, vaccine group differences in hearing persisted. Analysis stratified by the primary course head-to-head and combination schedules (the PREVIX_COMBO RCT) [5] suggested that slightly better hearing in the +P group was present across all primary schedules and was greatest in the SSSP+P versus SSSP+S comparison. The 19% higher prevalence of no hearing loss in the +P group (36%) compared to the +S group (16%) at age 18 months is also noteworthy as our surveillance over many years has consistently found less than 10% children are without OM at this age [3,20]. These unique findings are difficult to explain and should be interpreted with caution as not all randomised participants received hearing assessments.

A major limitation of our study is small sample size for this secondary outcome [10]. As mentioned, the low numbers receiving hearing tests was almost solely related to delays in implementation of hearing tests, rather than selection of participants based on certain characteristics or preferences. We show that baseline characteristics and risk factors of participants receiving hearing tests were not different between vaccine groups at any age. The high event rate somewhat compensates for low numbers and may increase confidence in the apparent vaccine group difference in hearing detected at primary endpoint age 18 months. Given the paucity of published data from vaccine trials to support or refute our findings, further investigation of biological plausibility is warranted, particularly as higher valency vaccines become available [21].

Our systematic review identified an almost complete absence of published contemporary (PCVera) data on the prevalence and persistence of hearing loss in very young First Nations children [22]. In addition, data from the NT hearing outreach services show that only 11% of services (either audiology, Clinical Nurse Specialist, or ear, nose, and throat (ENT) teleotology) were provided to children under age 3 years [23]. Of 1,741 First Nations children who received audiology services, 56 (3%) were under age 3 years and of these 17 (31%) had some level of hearing loss.

The systematic review [22] and national OM guidelines [14] also highlight the paucity of prevalence studies and absence of published intervention studies that address this crisis. In 2022 and 2023, the first community-based prevalence studies of hearing loss in First Nations children living in urban settings were published; one identified that at age 12 months, 69% of 67 children tested had hearing loss, including 24 (36%) with moderate hearing loss [24]. The second, larger study, which excluded children under age 3 years, obtained hearing data from 1,087 (median age 8·2 years) of 1,669 children enrolled; 279 (25·7%) had hearing loss [25]. Neither study was an evaluation of hearing loss rehabilitation or prevention strategies.

Using combined vaccine group data, our key findings identified for the first time the high prevalence of persistent and fluctuating mild to moderate hearing loss throughout early childhood, from 75% at age 12 months to 53% at age 36 months. In particular, the high prevalence of moderate hearing loss, defined by the World Health Organization as disabling for children [11], which was 42% at age 12 months, declining slowly to 19% at age 36 months.

These findings place many children in need of hearing assistance according to the Hearing Australia threshold of ≥23·3 dB [12]. Also, according to national OM guidelines [14], at ages 12 to 30 months, around 30% children met criteria for ENT consultation, based on the mean hearing level in the better ear of >30 dB.

Transitions in level of hearing loss between consecutive 6-monthly visits, despite stable prevalence rates at each time point, confirm the fluctuating nature of conductive hearing loss in this age group. This has practical implications that support the shift from screening programs to surveillance, commencing within first year of life [26].

A strength of our study is that all hearing tests were performed in sound-treated facilities by paediatric audiologists at scheduled study visits, regardless of clinical presentation. However, the majority of hearing tests were VROA for which the minimum test level varies according to test environment and child. If the minimum test level available was 20 to 25 dB, children with a true hearing level below test limits (i.e., no hearing loss) could be misidentified as having mild hearing loss using our study definitions, producing a false positive and overinflating the prevalence of mild hearing loss as determined by VROA. The prevalence of moderate (disabling) hearing loss is likely less susceptible to overinflation, and, at over 30% at every age until age 36 months, remains plausible and unacceptable.

From primary healthcare to specialist services, there are huge shortfalls that disrupt a child’s hearing care pathway [27]. At December 2022, there were 3,265 Indigenous children and young people on the wait list for fly-in/fly-out audiology outreach services and 2,284 on the ENT teleotology wait list [23]. As mentioned, in 2022, very few outreach services were for children under age 3 years, and ENT surgeries were rarely performed (4% of all ENT services were for grommets) [23].

This report of the extraordinarily high prevalence of preventable hearing loss in First Nations children highlights the need for effective social and public health reforms with greater investment in interventions that target social determinants, namely, more preventative models [28]. Others have demonstrated the inequalities in ENT service provision between Aboriginal and non-Aboriginal children, attributed to socioeconomic status and geographical isolation [29]. In the United States, social deprivation limits children’s access to appropriate medical and surgical treatment for their ear disease and increases odds of severe complications [30]. While ear and hearing problems are underserviced, effective prevention strategies are lacking and there is a dearth of efficacy studies that comprehensively investigate the benefits and harms of specialist ENT or audiological interventions for children living in remote regions [31]. Unfortunately studies of interventions that address social determinants will be difficult and few that report OM outcomes have been published or show a benefit [32]. Our research program is currently evaluating a primary healthcare workforce enhancement strategy (the Hearing for Learning Initiative) of remote on-country training and employment of ear health facilitators, which could provide a sustainable, culturally appropriate, and safe model to detect and prevent disabling hearing loss and support children who have hearing problems [33,34].

In conclusion, our discovery of a previously unrealised potential benefit of pneumococcal conjugate vaccination in reducing hearing loss should be investigated further as new expanded valency pneumococcal vaccines are developed. We strongly recommend more rigorous innovative trial designs, incorporating broader outcomes such as hearing loss and developmental milestones. This will improve understanding of the full impact of vaccination on all outcomes of importance, particularly those with downstream socioeconomic benefits for disadvantaged populations.

Ethics approval and consent to participate

Ethical approval has been obtained from Human Research Ethics Committees of the Northern Territory Department of Health and Menzies School of Health Research (NHMRC Reg no: EC00153), the Central Australian HREC (NHMRC Reg no: EC00155), and West Australian Aboriginal Health Ethics Committee (WAAHEC-377-12/2011). Parents or guardians provided signed informed consent for their infant’s participation.

Data safety monitoring

The study was overseen by an independent Data Safety and Monitoring Board (iDSMB).

Supporting information

Acknowledgments

We thank the many families living in remote communities in northern Australia who participated in the PREVIX_COMBO and PREVIX_BOOST randomised controlled trials and the PREVIX_VOICES study. We thank the many research nurses, audiologists, laboratory staff, and others who worked on the trials.