1 INTRODUCTION

Seasonal influenza is an infectious disease that highly affects population health in the Europe.1 Worldwide, annual influenza epidemics are estimated to result in about 3 to 5 million cases of severe illness, especially among older adults, young children (<5 years), pregnant women, and individuals with chronic medical conditions.2 In high-income countries, most influenza-related deaths occur among people aged 65 years or older.3 Respiratory complications are the most common sequelae,4 and it has been estimated that about 290,000 to 650,000 deaths from respiratory causes5, 6 and 99,000 to 200,000 deaths from lower respiratory tract infections (LRTIs) are attributable to influenza annually.7 Furthermore, several extra-respiratory complications, such as cardiovascular and nervous system play an important role.8 Worryingly, seasonal influenza generally represents an underappreciated public health problem with significant socio-economic implications.9

The monitoring and surveillance of seasonal influenza is possible through data collection and sharing systems, such as FluView in the United States (www.cdc.gov/flu/weekly) and FluNews in Europe (www.flunewseurope.org), that systematically collect data on seasonal influenza and publish periodic reports to inform on epidemiological trends. InfluNet is the Italian nationwide sentinel surveillance system for influenza, coordinated by the Italian National Institute of Health. It collects epidemiological (InfluNet-Epi) and virological (InfluNet-Vir) data that are weekly published on FluNews-Italy (https://www.epicentro.iss.it/influenza/FluNews) reports and uploaded into the European database coordinated by the European Centre for Disease Prevention and Control (ECDC). FluNews-Italy also integrates findings from other surveillance systems, such as the monitoring system of severe and complicated laboratory-confirmed cases of influenza, daily mortality among the elderly and InfluWeb (a web-based surveillance system of influenza-like illness [ILI]). More information on influenza burden (e.g., complications and hospitalizations) can be gathered from other sources, such as Health for All database (https://www.istat.it/it/archivio/14562) or published papers. However, the available evidence is still suboptimal. For instance, health technology assessment (HTA) projects of different preventive interventions against influenza have brought to light the need for more data.10-14

In sum, reliable country-specific data on influenza burden play a crucial role in informing the planning of prevention and control measures to limit the spread of the disease and minimize associated costs. For this reason, in scientific literature, there are some country specific reviews aiming to assess influenza incidence and clinical and economic burden. The published reviews focus either on a specific geographical area, such as Latin America, sub-Saharan Africa, Japan, or West Europe, or on a specific age range, such as elderly or pediatric age.15-21 To the best of our knowledge, no review has focused on Italy. Consequently, the present study aims to provide an overview of available data on the burden of seasonal influenza in Italy. Alongside the above-described Italian databases, this comprehensive review may be of aid for policy makers, health economists, public health practitioners, and other relevant stakeholders.

2 MATERIALS AND METHODS

This systematic review of the literature (PROSPERO registration number: CRD42021272644) was conducted following the 2020 PRISMA guidelines (Appendix A).

2.1 Identification of eligible studies

All studies quantifying the burden of seasonal influenza in Italy were potentially eligible, independently by initial influenza clinical presentation. The study outcome was the burden of influenza defined here as influenza-attributable complications, hospitalizations, or deaths. The study population was the entire Italian population, independently of age, health status, and any other variable affected by both laboratory-confirmed influenza and clinical proxies (e.g., ILI). No formal limits were established for study design. By contrast, the following exclusion criteria were applied: (i) studies evaluating the burden of pandemic influenza; (ii) case reports and case series with no possibility to establish the population burden of influenza; (iii) economic modeling with no original data; (iv) narrative reviews and other forms of the second-hand research; (v) original studies focusing only on epidemiological and/or virological surveillance of the laboratory-confirmed influenza and/or ILI.

The literature search was performed by consulting three databases, namely, PubMed, Web of Science (WoS), and Embase. The following search string was used on PubMed: “(epidemiology OR epidemiological OR virolog* OR surveillance OR incidence OR (“attack” AND rate) OR complicat* OR hospitalization OR (inpatient AND (admission OR care)) OR (outpatient AND (admission OR care)) OR (hospital AND (admission OR care OR discharge)) OR ambulatory OR mortality OR death OR sequelae OR visit) AND (influenza OR flu) AND (Italy OR italian)”; this spelling was then adapted to WoS and Embase. No search restrictions were applied. The search was updated to July 31, 2020.

After removing duplicates, papers were screened by title and abstract first. Clearly ineligible studies were discarded. The remaining records were assessed in the full-text modality.

2.2 Data extraction and synthesis

From the articles definitively included in the literature review, the following information were extracted: bibliographic record, study location, study setting (i.e., outpatient, inpatient, institutionalized, and mixed), main demographic characteristics of the study population (e.g., sample size, age, and sex distribution); study period/influenza season, type of outcomes, and their occurrence. Moreover, if available, data were stratified by viral (sub)type and type of outcome.

A meta-analysis of data was not planned because of the expected heterogeneity in study populations and endpoints. Indeed, data were summarized in a narrative way.

2.3 Quality assessment

The Newcastle-Ottawa Scale (NOS) was used for evaluating the quality of included studies. NOS adopts a star system, with a total score ranging from 0 to 9 and a score ≥7 indicating a high-quality study. Two investigators separately performed the quality evaluation of each study, and disagreements were settled by a joint re-evaluation of the original article with a third author. No study was excluded based on quality criteria.

3 RESULTS

The search of the three databases yielded a total of 9268 articles. After duplicates removal, 6640 articles were screened for title and abstract and 28 were selected for full text screening. It was not possible to retrieve four articles. Twenty-four articles were then screened by full text, and eight studies were excluded with the following reasons: did not meet the inclusion criteria (n = 2), not related to the topic (n = 4), reviews (=2). Eventually, 16 articles22-33 published from 2001 to 2020 were included in the qualitative synthesis. Details about the study selection process are shown in the flowchart (Figure 1).

Flow chart of the selection process

3.1 Studies' characteristics and quality

Seven studies (43.7%) were conducted in the Northern Italy,22, 23, 25, 27, 28, 30, 33 three (19.7%) in Central Italy,26, 31, 32 and one in Southern Italy.29 Four (25%) were nationwide studies24, 34-36 and one (6.2%) was conducted both in Lombardy (in Northern Italy) and Apulia (in Southern Italy).37

Five studies (31.2%) were set into a hospital,22, 25, 27, 32, 33 one (6.2%) in a residential drug-rehabilitation community,26 six (37.5%) in a primary care setting,23, 28-31, 37 and four (25%) were based on administrative health data.24, 34-36 Regarding studies set in a primary care setting, one (16.67%) was set into a health care unit,33 one (16.67%) in the district of a local health authority (LHA),30 and four (66.67%) based on general practitioners (GPs) or primary care pediatricians (PCPs).28, 29, 31, 37

Three articles (19.7%) concerned the pediatric age,27, 28, 32 three articles (18.7%) concerned adulthood,22, 26, 37 and four (25%) old age,25, 30, 31, 33 whereas four (25%) articles referred to the entire population without age distinction.24, 34-36

Eight studies reported data on only one season22, 26-29, 31, 33, 37; Rizzo et al reported data from 1970 to 2001,34, 35 and Bertolani et al reported data from 2008 to 2015.24 The other studies reported data on a period of up to four seasons.23, 24, 30, 32, 36 The characteristics of included studies are reported in Table 1.

TABLE 1. Studies' characteristics and data First author, year City Setting Study population Period/influenza season Mean age Male gender No. of participants Outcome (% calculated on the number of participants) No. deaths No. hospitalizations No. complications Bassetti M, 2019 Udine Hospital Patients admitted with laboratory-confirmed influenza 2017/18 48 (range 0–87) 51.7% 29 7 (24.1%) N.A. 19 (65.5%) Bellino S, 2019 Treviso Primary care (Local Health Unit) and hospital Three cohorts of elderly subjects 65+ 2014/15–2016/17 N.A. 43.5–44% 249,005 person-year (125,253 for vaccinated, and 123,752 for unvaccinated) 4855 and 3276 in vaccinated and unvaccinated subjects 18,355 (11,712 and 6643 in vaccinated and unvaccinated subjects) (ICD-9-CM 487, 480–486, 460–466, 490–496, 500–508, 510–516, 410, 422, 427, 428, in primary or secondary diagnosis) N.A. Boattini M, 2020 Torino, Lisbon, Nicosia (Cyprus) Hospitals Oldest-old (>85) patients admitted for laboratory-confirmed influenza and/or RSV infection or developing it during the course of admission for other causes 2017/18–2018/19 89.4 (±3.9 SD) 31.5% 251 35 (13.9%) N.A. 81 (32.27%) (radiological signs of pneumonia) Boschini A, 2006 Rimini Residential drug-rehabilitation community Former injecting drug users with ILI Feb–Mar 2004 N.A. 82.29% 209 N.A. N.A. 41 (19.62%) Bosis S, 2005 Milan Hospital Children <15 attending the emergency room Nov 1, 2002-Mar 31, 2003 4.0 (± 3.7 SD) 50.7% 223 (influenza positive) N.A. 12 (5.4%) 99 (44.39%) Esposito S, 2011 Northern Italy Primary care (PCPs) Healthy children <14 years of age without severe chronic medical condition but with signs and/or symptoms of ILI Nov 1, 2008-Apr 30, 2009 3.8 (± 2 SD) 51% 2143 (influenza positive) N.A. 16 (0.7%) 424 (19.78%) Loconsole D, 2019 Apulia region Primary care (sentinel-physician network of PCPs and GPs) and hospitals Patients with ILI, patients hospitalized with severe acute respiratory illness, patients admitted to all ICUs of the region with ARDS 2017/18 N.A. N.A. 565 (influenza positive) 23 (4.07%) (deaths occurred in patients with ICU-ARDS) 50 (8.85%) (ICU-ARDS hospitalizations) N.A. Mannino S, 2013 Cremona, Bergamo, Mantova, Lecco, Pavia Primary care (Local health authorities' district) and hospital Residents 65+ who sought influenza vaccination (excluding those in hospital, nursing homes or rehabilitation centers) 2006/07–2008/09 76.5 among aTIV; 74.9 among TIV 43.2% aTIV: 84,665 person-season, TIV: 79,589 person-season N.A. aTIV: 114 (0.135%), TIV:111 (0.139%) (ICD-9-CM 480–487) N.A. Manzoli L, 2009 Chieti, Pescara Primary Care (GPs) Elderly assisted by GPs First semester 2007 75.8 (± 7.4 SD) 43.4% 32,457 N.A. 142 (0.44%) (ICD-9-CM 480–487) N.A. Mastrolia M, 2019 Florence Hospital Children 1 month to 14 years with laboratory-confirmed influenza associated to neurological disease 2017/18–2018/19 27 months (IQR 7–48) 26.67% 15 N.A. N.A. 1 (6.7%) (impairment in motor skills) Mikulska M, 2013 Genova Hospital (HSCT unit) Adult outpatients seen at least once a month in the HSCT unit Jan 1 to Mar 31, 2011 N.A. 50% 20 (influenza positive) N.A. N.A. 1 (5%) (Clinical and radiological pneumonia) Sessa, 2001 Lombardy and Puglia Primary care (GPs) Patients visited for clinical influenza Dec 15, 1998, to Mar 15, 1999 40 49.9% 6057 N.A. 26 (0.43%) 2125 (35.1%) Bertolani A, 2018 Nationwide N.A. General population 2008, 2010–2015 N.A. N.A. N.A. N.A. Average annual number of hospitalizations: 17,488 (3508 observed 487 codes + 13,980 estimated from other codes) (ICD-9-CM 422, 427, 428, 460–466, 481–486, 487, 481–486, 490–496, 500–508, 510–516) N.A. Rizzo C, 2006 Nationwide N.A. General population 1970–2001 N.A. N.A. N.A. Excess mortality rate (per 100,000), P&I and AC respectively: 0.72 and 5.60 age 45–64; 14.13 and 98.86 age >65 (ICD-8 codes 480–486 and 470–474, ICD-9480–486 and 487) N.A. N.A. Rizzo C, 2007 Nationwide N.A. General population 1970–2001 N.A. N.A. N.A.

Excess deaths attributable to influenza: 57,234 from P&I, 318,806 from AC

Excess mortality rate (per 100,000), P&I and AC respectively (age-adjusted): 1.9–2.2 and 11.6–18.6 all ages; 0.4–0.7 and 4.3–6.6 age 45–64; 12.7–14.2 and 71.2–115.7 age 65 + (ICD-8 codes 480–486 and 470–474, ICD-9480–486 and 487)

N.A. N.A. Rosano, 2019 Nationwide N.A. General population 2013/14–2016/17 N.A. N.A. 5,290,000 estimated ILI

Excess deaths attributable to influenza: 68,068 (1.29% of ILI)

Excess mortality rate (per 100,000): 11.6–41.2 all ages; 65.0–147.3 (65+)

N.A. N.A. Abbreviations: AC, all cause; ARDS, acute respiratory distress syndrome; GP, general practitioner; HSCT, hematopoietic stem cell transplantation; ICU, intensive care unit; ILI, influenza-like illness; N.A., not available; PCP, primary care pediatrician; P&I, pneumonia and influenza; RSV, Respiratory Syncytial Virus; SD, standard deviation; TIV, trivalent inactivated vaccine; aTIV, adjuvanted trivalent inactivated vaccine.

The quality of studies varied in the range from 4 to 6 stars (median: 4; mean: 4.562) (Table 2). All the studies were judged to have a representative exposed cohort and a follow up long enough for outcomes to occur. The quality assessment was penalized by the absence of the non-exposed cohort that prevented assigning three stars for all the studies.

TABLE 2. Quality of included studies Author, year Representativeness of the exposed cohort Selection of the unexposed cohort from the same community as the exposed Ascertainment of exposure: secure record Outcome of interest not present at start of study Comparability of cohorts (on the basis of the outcome). Control for: Assessment of outcome Follow-up long enough for outcomes to occur Adequacy of follow up of cohorts Important factor Additional factors Bassetti., 2019 * N.A. * N.A. N.A. * * Bellino, 2019 * N.A. * N.A. N.A. * * Bertolani, 2018 * N.A. * N.A. N.A. * * Boattini, 2020 * N.A. * N.A. N.A. * * Boschini, 2006 * N.A. * N.A. N.A. * * Bosis, 2004 * N.A. * * N.A. N.A. * * * Esposito, 2011 * N.A. * * N.A. N.A. * * * Loconsole, 2019 * N.A. * N.A. N.A. * * Mannino, 2012 * N.A. * N.A. N.A. * * * Manzoli, 2009 * N.A. * * N.A. N.A. * * * Mastrolia, 2019 * N.A. * N.A. N.A. * * Mikulska, 2014 * N.A. * * N.A. N.A. * * * Rizzo, 2006 * N.A. * N.A. N.A. * * Rizzo, 2007 * N.A. * N.A. N.A. * * Rosano, 2019 * N.A. * N.A. N.A. * * Sessa, 2001 * N.A. * N.A. N.A. * * Note: Referring to the Newcastle Ottawa Scale, the star is allocated if methods adopted are considered acceptable. Abbreviation: N.A., not available. 3.2 Health burden of influenza 3.2.1 Complications

Eight articles (50%) evaluated influenza-related complications22, 25-28, 32,