Influenza vaccine effectiveness in immunocompromised patients with cancer: A Danish nationwide register-based cohort study

Abstract

Background: Influenza vaccination is free of charge for Danish citizens with acquired immunodeficiency but recommendations do not specifically target patients with cancer. This study investigated whether influenza vaccination reduces the main outcome of overall mortality and the secondary outcomes of influenza requiring treatment, pneumonia, myocardial infarction, stroke, heart failure, and venous thromboembolism in patients with cancer.

Methods: This was a register-based nationwide cohort study. Adjusted hazard ratios (aHRs) and 95% confidence intervals (95% CIs) for overall mortality and secondary outcomes were estimated using Cox proportional hazards models. Analyses were conducted separately for four subgroups: patients aged <65 years with solid tumors, patients aged ≥65 years with solid tumors, patients aged <65 years with hematological cancer, and patients aged ≥65 years with hematological cancer.

Results: A total of 53,249 adult patients with solid tumors who received chemotherapy and 22,182 adult patients with hematological cancer were followed for up to five influenza seasons in the study period of 2007-2018. In the main analysis covering December-March, influenza vaccination was associated with reduced overall mortality in all four subgroups. The reduction was most pronounced in patients with hematological cancer aged <65 years (aHR, 0.66; 95% CI, 0.51-0.87) and smallest in patients with solid tumors aged <65 years (aHR, 0.91; 95% CI, 0.84-0.99). In sensitivity analyses covering January-March, the aHR was 0.87 (95% CI, 0.65-1.16) in patients with hematological cancer aged <65 years and 1.01 (95% CI, 0.92-1.10) in patients with solid tumors aged <65 years. Results for the secondary outcomes were inconclusive.

Conclusions: The results of this study cannot reject that influenza vaccination reduces overall mortality in immunocompromised patients with cancer. The results must be interpreted with caution because of potential unmeasured confounding, which can result in the overestimation of influenza vaccine effectiveness.

Keywords: chemotherapy; hematological cancer; immunosuppression; influenza vaccination; seasonal influenza; solid tumors.

FIGURE 1

Flowchart of the study population selection. ^aDiagnosis codes are available in Appendix A1. ^bAge in the first observed influenza season for each patient.

FIGURE 2

Number of events, incidence rates per 10,000 person‐years, and hazard ratios for the associations between influenza vaccination and outcomes during the influenza season (December–March) for all seasons combined (2007/2008–2017/2018). Incidence rates are shown as events per 10,000 person‐years. Hazard ratios from Cox proportional hazards models, with days measured from December 1 in each influenza season as the underlying time. The adjusted models include influenza vaccination in the previous season, criteria that qualify for free‐of‐charge influenza vaccination, education level, place of birth, age, and sex. Cancer stage at diagnosis was included in the adjusted analyses of patients with solid tumors, and receipt of chemotherapy was included in the adjusted analyses of patients with hematological cancer. CI indicates confidence interval; HR, hazard ratio; IR, incidence rate.

FIGURE 3

Sensitivity analysis of influenza seasons with a match between the influenza vaccine components and circulating influenza virus and only including follow‐up during the active influenza season defined as the proportion of influenza‐like illness consultations >0.75%. Hazard ratios from a Cox proportional hazards model adjusted for influenza vaccination in the previous season, criteria that qualify for free‐of‐charge influenza vaccination, education level, place of birth, age, and sex. Cancer stage at diagnosis was included in the analysis of patients with solid tumors, and receipt of chemotherapy was included in the analysis of patients with hematological cancer. The underlying time is days measured from the date when the proportion of influenza‐like illness consultations first was >0.75%. Appendix A7 specifies the included seasons and time periods. CI indicates confidence interval; HR, hazard ratio.

FIGURE 4

Associations between influenza vaccination and overall mortality across the year divided into 1‐month periods. Hazard ratios from a Cox proportional hazards model adjusted for influenza vaccination in the previous season, comorbidities resulting in free‐of‐charge influenza vaccination, education level, place of birth, age, and sex. Furthermore, cancer stage was included in the analysis of patients with solid tumors, and chemotherapy was included in the analysis of patients with hematological cancer. CI indicates confidence interval; HR, hazard ratio.

FIGURE 5

Adjusted hazard ratios for the associations between influenza vaccination and overall mortality observed in the modified periods of January–March and January–April for all seasons combined (2007/2008–2017/2018). Hazard ratios from a Cox proportional hazards model adjusted for influenza vaccination in the previous season, criteria that qualify for free‐of‐charge influenza vaccination, education level, place of birth, age, and sex. Cancer stage at diagnosis was included in the analysis of patients with solid tumors, and receipt of chemotherapy was included in the analysis of patients with hematological cancer. The underlying time is days measured from January 1 in each influenza season. CI indicates confidence interval; HR, hazard ratio.