Prevalence of COVID-19-like symptoms in Italy and Lombardy, March-April 2020, and their implications on cancer prevention, diagnosis and management.

Two population-based surveys on coronavirus disease-2019 (Covid-19)-like symptoms were conducted by BVA-Doxa on representative samples of the general population from Italy and its largest region, Lombardy, with over 10 million inhabitants and heavily struck by Covid-19, on 27-30 March and 3-7 April, on a total of 2000 individuals - 353 from Lombardy. Overall, 14.2% of Italians - and 19.6% in Lombardy - reported Covid-19-like symptoms. The estimated prevalence was higher in the young, in smokers, and in Lombardy over the period 20 March to 7 April. Although the influenza season was ended by mid-March, at least part of the symptoms may be not Covid-19 related. Even assuming that only half are, at least 7% of Italians and 10% in Lombardy had been affected by Covid-19. To these, asymptomatic or paucisymptomatic cases have to be added. These estimates are at least one or two orders of magnitude larger than official registered cases. This has major implications for cancer prevention, management and treatment.

Several thousand patients infected with coronavirus disease-2019 (COVID-19) were projected in Italy by mid-March (Remuzzi and Remuzzi, 2020), and registered COVID-19 cases were over 110 000 by the end of March (Boccia ).

Italy, and its largest region, Lombardy (with 10 million inhabitants), have been also subsequently heavily hit by the COVID-19 epidemic. According to official data (Protezione Civile, 2020), up to 21 April 183 857 cases and 24 648 deaths were registered in Italy. Corresponding figures for Lombardy were over 67 931 cases and over 12 579 deaths.

The COVID pandemic has substantial health implications besides its direct effects on population morbidity and mortality (Rosenbaum, 2020). Cancer screening, diagnosis and management have been heavily affected by the current pandemic, and it is, therefore, important to provide figures on Covid cases and deaths, in order to provide a perspective for the evolution of the pandemic and its potential implications on cancer patients (Nelson, 2020; Vrdoljak ). Each year, in Italy, about 180 000 cancer deaths are registered, corresponding to over 350 000 incident and 2 million prevalent cases (Santucci ). These figures have to be put in perspective with the current Covid cases, deaths and hospital burden.

The real number of Covid cases is; however, undefined, and in Italy registered cases include essentially hospital admissions, plus a small number of positive subjects from nonorganized PCR tampons.

To obtain an estimate of possible number of cases, we included a series of questions on COVID-19-related symptoms (including fever, headache, cold, cough, gastrointestinal disorders-plus, separately, fever over 38.5°C), in two weekly national surveys routinely conducted by BVA-Doxa. The first survey was conducted on 27–30 March, and based on a sample of 1000 individuals, representative of the general Italian population aged 18–85 in terms of sex, age, geographic area and socioeconomic status. Of these, 169 subjects were resident in Lombardy. The second survey was conducted on 3–7 April, using a different sample identified with the same criteria also included 1000 subjects, 184 from Lombardy. Thus, the two waves combined cover the period 7 March–7 April. The data were collected using computer-assisted web interviews and related to the 3-week preceding data collection. Confidence intervals of percentage estimates were based on the binomial distribution. We also computed the odds ratios (ORs), and the corresponding 95% confidence intervals (CIs) of reporting Covid-19-like symptoms, including terms for sex, age, smoking, education, wave of interview, and area of residence.

The key findings are given in Table 1. In Italy, 14.2% of subjects reported some COVID-19-like symptoms, and 1.3% fever over 38.5°C over the last 3 weeks. Corresponding figures for Lombardy were 19.6% for any symptom, and 2.3% for fever over 38.5°C. Both in Italy and in Lombardy, the proportion of subjects reporting COVID-19-like symptoms was higher in women than in men, at younger age than in middle age and the elderly, in smokers than in nonsmokers, and in subjects with higher education. For age and smoking, the differences were significant.

Table 1

Prevalence (%) and 95% confidence intervals of Covid19-like symptoms in the last three weeks in a representative sample of the Italian population. BVA-Doxa, Italian national surveys, 27–30 March and 3–7 April 2020

Table 2 gives the corresponding multivariate ORS. These were 2.0 (95% CI, 1.5–2.8) for the young vs. the elderly, 1.6 (95% CI, 1.2–2.1) for smokers vs. nonsmokers, and 1.7 (95% CI, 1.2–2.3) for Lombardy ves other Italian regions.

Table 2

Prevalence odds ratio and 95% confidence intervals of Covid-19-like symptoms in the last 3 weeks in a representative sample of the Italian population. BVA-Doxa, Italian national survey 27–30 March and 3–7 April 2020

At least part of the symptoms reported is not COVID-19 related. Although the annual flu season had ended by mid-March – and was particularly moderate the last winter, part of the symptoms reported may well be related to other, nonspecific (viral) conditions. It is possible; however, that a considerable proportion of symptoms – and most of fever over 38.5°C in March and early April – were due to COVID-19. Even assuming that only about half of reported symptoms were due to COVID-19, over 7% of the population in Italy and 10% in Lombardy would have been affected by COVID-19 in the weeks covered by data collection only. This would correspond to at least 5 million subjects affected in Italy, and 1 million in Lombardy alone, a figure that can be doubled assuming that most COVID-19-like symptoms are indeed COVID-19 related. The information was restricted to a 2-week period, and additional subjects likely had been affected before 20 March. In addition, an undefined proportion of the population has been affected by totally asymptomatic or minimally symptomatic, COVID-19, and was not reported in our surveys.

While the key limitation of the surveys considered are their subjective, self-reporting of symptoms, they have the strength of being included in a routinely weekly conducted, validated survey, of being reasonably large, and representative of the general population of Italy and Lombardy.

The larger proportion of symptoms at younger age likely reflects more active social interactions in this age group.

These data indicate therefore that – even ignoring asymptomatic cases – the COVID-19 epidemic may well have affected a substantial proportion of the Italian population by early April, certainly at least one order of magnitude larger, and possibly up to two orders of magnitude larger, than the registered cases.

Still, these estimates remain far from those required to reach adequate natural immunity on a population level and indicate therefore that the COVID pandemic is going to persist in Italy at least for the next few months.

This would, therefore, imply unique challenges to the diagnosis and management of other diseases, including cancer.

Acknowledgements

The survey was conducted and supported by BVA-Doxa with internal funding.

Conflicts of interest

There are no conflicts of interest.