Background: UK specific data on the risk of developing hospitalised CAP for patients with underlying comorbidities is lacking. This study compared the likelihood of hospitalised all-cause community acquired pneumonia (CAP) in patients with certain high-risk comorbidities and a comparator group with no known risk factors for pneumococcal disease.
Methods: This retrospective cohort study interrogated data in the Hospital Episodes Statistics (HES) dataset between financial years 2012/13 and 2016/17. In total 3,078,623 patients in England (aged ≥18 years) were linked to their hospitalisation records. This included 2,950,910 individuals with defined risk groups and a comparator group of 127,713 people who had undergone tooth extraction with none of the risk group diagnoses. Risk groups studied were chronic respiratory disease (CRD), chronic heart disease (CHD), chronic liver disease (CLD), chronic kidney disease (CKD), diabetes (DM) and post bone marrow transplant (BMT). The patients were tracked forward from year 0 (2012/13) to Year 3 (2016/17) and all diagnoses of hospitalised CAP were recorded. A Logistic regression model compared odds of developing hospitalised CAP for patients in risk groups compared to healthy controls. The model was simultaneously adjusted for age, sex, strategic heath authority (SHA), index of multiple deprivation (IMD), ethnicity, and comorbidity. To account for differing comorbidity profiles between populations the Charlson Comorbidity Index (CCI) was applied. The model estimated odds ratios (OR) with 95% confidence intervals of developing hospitalised CAP for each specified clinical risk group.
Results: Patients within all the risk groups studied were more likely to develop hospitalised CAP than patients in the comparator group. The odds ratios varied between underlying conditions ranging from 1.18 (95% CI 1.13, 1.23) for those with DM to 5.48 (95% CI 5.28, 5.70) for those with CRD.
Conclusions: Individuals with any of 6 pre-defined underlying comorbidities are at significantly increased risk of developing hospitalised CAP compared to those with no underlying comorbid condition. Since the likelihood varies by risk group it should be possible to target patients with each of these underlying comorbidities with the most appropriate preventative measures, including immunisations.