The INfluenza Vaccine Indication During therapy with Immune checkpoint inhibitors (INVIDIa-2) study was a multicenter prospective observational trial. The primary objective was to investigate the effectiveness of influenza vaccine administration in terms of incidence and severity of ILI in patients with advanced cancer undergoing systemic treatment with ICIs. The secondary objectives of the study regarded influenza vaccine safety during immunotherapy and the oncological outcome of patients to ICIs basing on the vaccine administration or influenza occurrence.

All patients with advanced solid tumors receiving therapy with ICI (alone or in combinations) from October 1, 2019, to January 31, 2020 (corresponding to the influenza vaccinal season) were eligible for enrollment. ILI, according to ECDC, was defined as a sudden and rapid onset of one or more of the following general symptoms: fever or low-grade fever, malaise and/or fatigue, headache, myalgia; and at least one of the following respiratory symptoms: cough, sore throat, dyspnea. “Sudden and rapid” is taken to mean to the extent that the date of symptom onset can be clearly identified.3 4

Since placebo-controlled randomized controlled trials of antiviral vaccination among adults with cancer are often considered ethically questionable19 and foreseeing a proportional numeric distribution due to the relatively low vaccine coverage expected in this population,12 we opted for an observational approach. We prespecified a comparative analysis of clinical characteristics between the two groups of vaccinated and unvaccinated patients to measure the indication bias deriving from the study’s observational nature.

The study’s primary endpoint was the time-adjusted ILI incidence, calculated in terms of time-to-ILI (TTI), from October 1, 2020, until April 30, 2020. Such a time-adjusted endpoint was chosen to weight the rate of ILI occurrence based on the time exposure of patients to the influenza season, considering the possibility of late enrollment or early death during the predefined period of the ILI observation. TTI was defined as the time interval between the date of enrollment, and the date of ILI symptoms occurrence; patients not experiencing ILI at the cut-off date of April 30, 2020, were considered as censored at this date or the date of death if previously occurred.

The secondary endpoints regarding ILI severity were the incidence of ILI-related complications, defined as pneumonia confirmed by imaging, laboratory-confirmed bacterial superinfections, or other severe complications; the ILI-related intravenous therapy administration; the ILI-related hospitalization rate; the duration of ILI symptoms (in days), the ILI lethality. Vaccination safety was recorded as a secondary endpoint in terms of vaccine-related adverse events (VRAEs), defined by the suspicion of a causal relationship between the influenza vaccine and the adverse reaction, and classified by Common Terminology Criteria for Adverse Events (CTCAE) V.5.0.20

Secondary endpoints regarding the patients’ oncological outcome to ICI treatment will also be measured in future analysis in terms of ICI safety, activity and efficacy (follow-up still ongoing as per protocol).

The influenza season, from October 2019 to April 2020, constituted the ILI observation period for both the primary and secondary endpoints regarding the severity of ILI. The timing of the study is shown in figure 1A.

Prespecified subgroup analyses were planned for elderly and patients with lung cancer, as populations at high-risk of ILI-related complications. Two prespecified age cut-offs were used for operational purpose: 65 years, which is the age to recommend influenza vaccination in Italy, and the median age of the study population, to consider equal size subgroups. The type of influenza vaccine received, the vaccination date, the figure recommending vaccination, and information about any other vaccine administered after the study enrollment were recorded, together with clinical characteristics and outcomes of patients about ILI, other viral or bacterial infections, and cancer disease and therapy. The study data were prospectively collected in a validated eCRF platform.

The study’s funders had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had final responsibility for the publication.

Based on an expected ILI incidence of 12% for unvaccinated ICI-treated patients with cancer,15 establishing a 50% reduction in the incidence of ILI in vaccinated patients as the reliable objective of the vaccination, statistical power of 80% and statistical significance with p<0.05, and hypothesizing vaccine coverage of 75% (based on WHO recommendation for categories at high risk21), an overall study sample of 974 patients was calculated. An overaccrual of 30% was preplanned based on the expected rate of screening failures, considering the trial’s observational nature. The study involved 82 Italian centers.

Data related to quantitative variables were reported using median and IQR and absolute count and percentages when referring to categorical items. The exact 95% CIs were calculated for proportions. Association between categorical variables was evaluated with the χ² test. TTI curves were estimated by the Kaplan-Meier method and compared with the log-rank test. The Cox proportional hazard model was used to assess associations between demographic and clinical factors and derive HRs and 95% CI. A logistic regression model was used when evaluating ILI complications, and OR with their 95% CI were reported. All the statistical tests were two tailed.

The INVIDIa-2 study prospectively enrolled 1279 patients with advanced cancer receiving ICIs. Of them, 1188 were eligible for the primary analysis (patient disposition is showed in figure 1B). The characteristics of the study population are summarized in table 1.

Overall, 581 patients received influenza vaccination (48.9%) while 607 did not (51.1%). Vaccinated patients were significantly more frequently elderly (p<0.0001), males (p=0.004), with poor European Cooperative Oncology Group (ECOG) performance status (PS, p=0.009), affected by lung cancer (p=0.01), and by other non-cancer comorbidities (p<0.0001) when compared with unvaccinated (table 1). Indication for influenza vaccination was given by general practitioners in 64.1% of cases, clinical oncologists in 16.0%, unsolicited patients’ decision in 12.5%, other physicians in 7.4%. The vaccine types administered in the study population are showed in online supplemental table S1. Patients received other vaccines during the study conduct: 45 received pneumococcal vaccination (43 of whom together with influenza vaccine), two patients received tetanus vaccination.

The incidence of ILI in the overall study population was 98 cases (8.2%, 95% CI 6.7 to 10.0). The ILI clinical manifestation in these patients, considering the ECDC definition, is represented in figure 2. The median duration of ILI symptoms was 7 days (IQR 4–14). The overall complication rate was 24.5%, with pneumonia developed during ILI in 21 cases (21.4%) and other complications in 3 (3.1%); intravenous therapies due to ILI were administered in 20.4% of all ILI cases, and ILI-related hospital admission was required in 11.2% of cases.

The time-adjusted ILI incidence, in terms of TTI, was similar in vaccinated and unvaccinated patients (figure 3, p=0.62). The crude incidence of ILI was 8.8% (95% CI 6.6 to 11.4) in vaccinated patients (51/581) versus 7.7%–95% CI 5.7 to 10.2 in unvaccinated patients (47/607) (p=0.52). Also, the median duration of ILI symptoms was similar in vaccinated and unvaccinated subjects (7 days and 8.5 days, respectively, p=0.82).

ILI complications were significantly less frequent for patients receiving vaccination than for unvaccinated patients: respectively, the complication rate was 11.8% (6/51) versus 38.3% (18/47), p=0.002 (figure 4). The type of complications found in the two groups is summarized in online supplemental table S2. ILI-related intravenous therapies were also significantly less frequent in vaccinated patients than in unvaccinated: 11.8% versus 29.8% (p=0.027).

Hospital admission due to ILI occurred for 11 patients (11.2% of all ILI cases); of them, 7 were unvaccinated (ILI-related hospitalization was required for 7/47 (14.9%) of unvaccinated versus 4/51 (7.8%) of vaccinated patients, p=0.27).

ILI lethality was 2% overall (2/98 patients, both unvaccinated). Within vaccine groups, ILI lethality was 0/51 (0%) in vaccinated versus 2/47 (4.3%) in unvaccinated patients.

Considering the respective timings of influenza vaccination and ILI occurrence, 4 patients (4.1%) developed ILI within 2 weeks after their vaccination, and 5 (5.1%) had the vaccine after a prior episode of ILI. Overall, these nine patients were likely not protected by their vaccination: nevertheless, the study results remained unchanged when removing these cases from the analyses (data not shown).

The relative vaccine effectiveness of adjuvanted versus non-adjuvanted vaccines in preventing ILI was 54.9% (6.3% vs 9.7%); this difference however was not statistically significant (p=0.21).

A Cox regression model was implemented to evaluate associations between demographic and clinical characteristics and TTI. The following were considered: sex, age, ECOG-PS, comorbidities (respiratory, cardiovascular or metabolic comorbidity), lung cancer as the primary tumor, influenza vaccination. After a selection based on p values, lung cancer (p=0.009), respiratory comorbidities (p=0.007) and age <69 years (p=0.028) resulted as independent factors associated with higher ILI incidence (table 2). When considering only vaccinated patients (n=581), none of the considered factors were associated with ILI incidence (online supplemental table S3).

When considering patients developing ILI (n=98), the only factor inversely associated with ILI-related complications at the multivariate analysis was the vaccine administration (p=0.004, table 2).

Overall, 9 of 581 patients had VRAEs: local erythema, local pain, local edema, fever, heart-pounding, rash and pruritus were reported. The overall incidence of VRAEs was 1.5%, with grade (G)1–2 according to CTCAE 5.0, self-limiting course and complete resolution; no G3-4 VRAEs were reported.

Overall, elderly patients (≥69 years) had non-statistically significantly longer TTI (p=0.15) than non-elderly. Patients affected by lung cancer had statistically significantly shorter TTI than patients with other primary tumors (online supplemental figure S4). The crude incidence of ILI in patients with lung cancer was also higher, 10.1% vs 6.1% in patients with non-lung tumors, respectively (p=0.013).

Considering the impact of influenza vaccination in these subgroups, it was consistent with that on the overall population: ILI incidence was similar irrespective of vaccination (online supplemental figure S5), and the rates of complications were lower in vaccinated patients when compared with those of unvaccinated individuals (online supplemental table S6).