Discussion

This ten-year review of pneumococcal BSIs identified opportunities for improvement in clinical practice. With regard to Antimicrobial Stewardship (AMS), it was uncommon for patients to be de-escalated from empiric broad spectrum empiric agents to narrow spectrum penicillins, despite availability of antimicrobial susceptibility results to support such decisions. This limited de-escalation to narrow-spectrum β-lactams suggests that therapeutic decisions were often guided by routine clinical practice or clinician preference for broad-spectrum coverage rather than tailored to microbiological confirmed data. ^{Reference Heenen, Jacobs and Vincent5} Targeted prescriber education, coupled with structured AMS support may facilitate more appropriate de-escalation of therapy. ^{Reference Heenen, Jacobs and Vincent5} Additionally, embedding automatic AMS review of positive blood cultures, incorporating electronic prescribing prompts at the time of result verification, and standardizing bacteremia management pathways may enhance microbiology-directed optimization of therapy and strengthen AMS impact. ^{Reference Schuts, Hulscher and Mouton6} On a more positive note, the proportion of IVOST was high, with over three-quarters of patients transitioned to PO therapy after a short median IV duration of 3.5 days.

The observed overall mortality rate of 6% was lower than historical estimates of IPD mortality among older adults, which have typically ranged between 10%–20% in international cohorts. ^{Reference van der Poll and Opal7} This may reflect improvements in early clinical recognition, timely commencement of antimicrobial therapy, and advances in supportive care. However, ICU admission in 10% of cases underscores the clinical severity associated with IPD.

The decline in pneumococcal BSIs during 2020–2021 coincided with the COVID-19 pandemic, likely reflecting the indirect effects of non-pharmaceutical interventions, including reduced social contact, widespread use of face masks or coverings and mobility restrictions, which were associated with marked declines in respiratory pathogen transmission. ^{Reference Brueggemann, Jansen van Rensburg and Shaw8} Contemporaneous national laboratory surveillance data in Ireland also recorded a reduction in typed isolates from patients with IPD from 374 (2019) to 181 (2020) and 160 (2021), followed by a resurgence to 379 (2023). ³ In addition, changes in healthcare utilization during the pandemic, including reduced emergency department attendance for non-COVID respiratory illness and modified hospital admission practices, may have influenced case ascertainment. ^{Reference Moynihan, Sanders and Michaleff9} Comparable reductions in invasive bacterial disease, including S. pneumoniae, were reported across multiple countries during periods of strict COVID-19 control measures; ^{Reference Brueggemann, Jansen van Rensburg and Shaw8} however, the subsequent return to prepandemic case numbers underscores that these reductions were transient and mainly reflective of behavioral and public health interventions rather than a sustained change in pneumococcal epidemiology.

Consequently, prevention of IPD remains a public health priority, particularly among older adults who bear the highest disease burden in Ireland, with an incidence rate of 24.7 per 100,000 in ≥65 years in 2023. ³ PCV and PPV are available as part of national immunization strategies. Despite this, prior pneumococcal vaccination was documented in only 31% of patients in our cohort. Furthermore, despite pneumococcal BSI being a clear indication for immunization, a vaccination recommendation at discharge was recorded in just 22% of cases, and only 26 of 38 recommendations were subsequently actioned. These findings highlight missed opportunities for secondary prevention.

Ireland introduced PCV7 into the childhood immunization schedule in 2008, replaced by PCV13 in 2010, resulting in substantial reductions in vaccine-type IPD and indirect protection in older age groups through herd immunity. ¹⁰ However, uptake of three PCV doses by 24-months has declined from 92.5% in 2015 to approximately 82% nationally in recent years. ^10,11 The median age in our cohort was 69.5 years, indicating that most patients would not have benefited from routine childhood PCV immunization and instead relied on adult vaccination strategies. PPV23 is recommended in Ireland for adults aged ≥65 years and for younger adults with defined risk factors. Since 2015, a sequential strategy of PCV13 followed by PPV23 has been recommended for adults with immunocompromising conditions or significant comorbidity, reflecting efforts to broaden protection in high-risk groups. ¹⁰ However, adult uptake has historically been low, 27%–36%, and adult vaccination often depends on patient health-seeking behavior, opportunistic primary care engagement, and clinician recommendation. ¹¹ Vaccine hesitancy, competing health priorities, and lack of systematic recall mechanisms may further contribute to suboptimal uptake. ^{Reference Williams12}

The high prevalence of chronic diseases (heart, kidney and pulmonary) and diabetes mellitus in this cohort suggests that many patients would likely have been vaccine candidates prior to their IPD case. The low rates of documented vaccination and discharge recommendation suggest a need for system-level interventions. Embedding vaccination assessment into standardized discharge protocols, incorporating electronic prompts, and strengthening communication with family physicians may improve vaccine uptake. Education of both healthcare professionals and patients regarding the importance of pneumococcal vaccination following IPD is also warranted. ^{Reference Cataldi, Kerns and O.’Leary13}

Serotype distribution in this study further informs vaccine strategy. While only 25% of cases were attributable to PCV13 serotypes, 70% would be covered by PCV20, and 77% by PPV23. These findings support consideration of higher-valency conjugate vaccines in immunization programs and reinforce the importance of continued surveillance to guide national vaccine policy. ^10,14 Given Ireland’s aging population and the burden of comorbidity observed, systematic assessment of pneumococcal vaccination status in community-based care settings, as well as during hospital encounters should be prioritized.

This study provides information about pneumococcal BSI over a ten-year period, but there are several limitations. Most BSI (78%) were attributed to a respiratory source, indicative of IPD most commonly arising in the context of pneumonia. The focus of this study was laboratory-confirmed BSI as a robust and microbiologically-defined end point. Non-bacteremic pneumococcal infections, including community-acquired pneumonia were not captured and likely represent a substantial additional burden of disease. Nonetheless, the observed patterns in serotype distribution, vaccination gaps, and antimicrobial management in IPD have clear implications for the broader clinical spectrum of pneumococcal infection. The retrospective design inherently relies on the completeness and accuracy of existing medical records and laboratory data, which may have led to under-reporting of complications, vaccination documentation, notably oral advice provided to patients regarding vaccination was not captured, or treatment decisions. Secondly, pneumococcal serotyping was not available for 7% of cases, potentially limiting the accuracy of vaccine coverage estimates. Thirdly, the absence of broader hospital denominator data precludes assessment of testing rates or prescribing behaviors across all patients with suspected pneumococcal disease. In addition, HIV serostatus and documented allergy status were not routinely available in the data set; both may have influenced clinical decisions around suitability for de-escalation from a cephalosporin to a penicillin for example. Furthermore, information on certain recognized risk factors for IPD—including alcohol misuse, asplenia or hyposplenism, and other forms of immunosuppression—were not consistently available, and the burden of comorbidity in this cohort is therefore likely underestimated. Finally, data from a single institution may limit generalizability, although findings are consistent with broader national trends and highlight systemic gaps in diagnostic and preventive practices.