The association between influenza vaccination and hospitalization rates of respiratory and cardiovascular conditions among elderly members of a private medical scheme during the winter season of 2004 in South Africa

Abstract

Objective: The objective of this nested case-control study was to investigate the association between influenza vaccination and hospitalisation rates for acute respiratory illness or cardiovascular disease and death from any cause among elderly population of a medical scheme in South Africa during the winter season of 2004. The study also reported on further sub-groups with different risk profiles within this elderly population. Method: All data were obtained from a computerised database containing information on a cohort of 45 700 elderly members. Cases were identified from hospital records and membership data between May and July 2004, and for each case four controls were chosen randomly from the cohort. Data on baseline characteristics were compared between cases and controls. Variables that may be potential confounders were identified and multivariate conditional logistic regression analysis was carried out to establish the vaccine effectiveness independent of confounding factors for the combined endpoints as well as for each endpoint separately. Analysis was done to confirm potential effect modification. Patients were classified as high-risk or low-risk according to their chronic illness profiles. Sub-group analysis was performed to determine the vaccine effectiveness in specific sub-groups of elderly patients. Results: In total 1 282 (2.81%) elderly had an outcome; 627 died from any cause, and 428 and 227 were hospitalised for cardiovascular conditions and respiratory conditions respectively. Analysis showed a 19.3% (95% CI, 3.1 to 32.9) reduction in the incidence of combined outcomes in the vaccinated group. All-cause mortality was reduced significantly by 23.6% (95%CI, 1.0 to 41.0) but there was no statistically significant reduction in either respiratory or cardiovascular hospitalisations. With statistical analysis, presence of chronic respiratory disease was identified as a potential effect modifier. Sub-group analysis was performed and after adjusting for confounding, vaccine effectiveness was determined as 56,3% (95% CI, 19.2 to 76.3) for all-cause mortality, 24.7% (95% CI, -37.7 to 58.8) for respiratory hospitalisation, and 44.4% (95% CI, -7.5 to 71.2) for cardiovascular hospitalisation in the sub-group of patients with chronic respiratory disease. After adjusting for confounding, vaccination was associated with a statistically significant reduction in all outcomes of 19.9% (95% CI, 2.2 to 34.4) in the high-risk sub-group of elderly, while in the low-risk sub-group of elderly there was a non-significant increase of 48.4% (95% CI, -289.7 to 43.4 ) in the outcomes. Conclusion: In this population of elderly influenza vaccination was associated with a statistically significant reduction in the combined outcomes. Estimates of effectiveness differ substantially between different sub-groups of high- and low-risk patients implying a possible greater benefit from vaccination in the high-risk elderly population.