Skip to main content
NCBI home page
BMC Public Health logoLink to BMC Public Health
. 2021 Nov 11;21:2069. doi: 10.1186/s12889-021-12126-4

The association between first and second wave COVID-19 mortality in Italy

Marco Vinceti 1,2,✉, Tommaso Filippini 1, Kenneth J Rothman 2,3, Silvia Di Federico 1, Nicola Orsini 4
PMCID: PMC8582237  PMID: 34763690

Abstract

Background

The relation between the magnitude of successive waves of the COVID-19 outbreak within the same communities could be useful in predicting the scope of new outbreaks.

Methods

We investigated the extent to which COVID-19 mortality in Italy during the second wave was related to first wave mortality within the same provinces. We compared data on province-specific COVID-19 2020 mortality in two time periods, corresponding to the first wave (February 24–June 30, 2020) and to the second wave (September 1–December 31, 2020), using cubic spline regression.

Results

For provinces with the lowest crude mortality rate in the first wave (February–June), i.e. < 22 cases/100,000/month, mortality in the second wave (September–December) was positively associated with mortality during the first wave. In provinces with mortality greater than 22/100,000/month during the first wave, higher mortality in the first wave was associated with a lower second wave mortality. Results were similar when the analysis was censored at October 2020, before the implementation of region-specific measures against the outbreak. Neither vaccination nor variant spread had any role during the study period.

Conclusions

These findings indicate that provinces with the most severe initial COVID-19 outbreaks, as assessed through mortality data, faced milder second waves.

Keywords: COVID-19, Epidemiology, Mortality, Public health, SARS-CoV-2, Waves

Background

The COVID-19 pandemic in many countries has been characterized by waves of infection. These recurrent local exacerbations of the pandemic present an opportunity to study the spread of the virus within a population. Key features of the pandemic are still not well understood, such as the susceptibility of the population to subsequent waves after the first outbreak, the threshold for herd immunity, the role of superspreaders [1–7] as well as of meteorological and environmental factors [8–11].

We compared the two COVID-19 waves within Italy, where the geographical distribution of the SARS-CoV-2 infection spread and the COVID-19 incidence was uneven during the first wave of the pandemic [12]. From an analysis within provinces of the official case counts during the first and second wave of SARS-CoV-2 infection, we found evidence of an association in the occurrence between the two periods. When incidence in the first wave was low (< 500 cases/100,000/day), the second wave incidence tended to be positively correlated, whereas a high first-wave incidence was strongly and inversely correlated with the second wave incidence. This observation raised the possibility that greater spread at the beginning of the pandemic could have induced some protection at the population level, resulting in a milder second wave, despite low levels (10–20%) of anti-SARS-CoV-2 antibody seroprevalence even in the areas most severely hit.

One problem with our earlier finding is that the incidence of SARS-CoV-2 infection is difficult to measure. It depends on the severity of the clinical course, the availability of molecular tests and the policy underlying their use at the population level. In Italy, for instance, the implementation of nasopharyngeal swabs and molecular tests to detect SARS-CoV-2 infection in the first wave was restricted primarily to symptomatic patients with suspected COVID-19. In the summer of 2020, however, the public-health policy was modified to extend testing to contacts of cases and to individuals without clinical symptoms, nearly tripling the number of daily molecular tests from the first to the second wave time periods [13]. An additional change of policy, starting on November 6, 2020, during the second wave, was the mandatory adoption of differential area-specific public-health measures [14, 15].

The area-specific relation between the two waves may be better assessed by a more objective measures of COVID-19 spread such as mortality rate, instead of incidence rates that are subject to swab testing policy and availability. Recently, COVID-19 death data for Italy in 2020 have become available with greater geographic detail, with figures available for the 107 provinces rather than the 20 larger regions within Italy. We used these mortality data to revisit the relation of the dynamics of the first and second waves of the outbreak.

Methods

Study population and outcome

We used data made freely available from public sources. We accessed 2020 monthly province-specific COVID-19 deaths as reported by the Italian National Institute of Statistics – ISTAT [16]. Provinces are the administrative entities that are intermediate between smaller municipalities and larger regions within Italy. Within 2020, we examined two time frames: we considered the first wave to be the period from the pandemic onset (February 24) to June 30, when the number of new infection cases had substantially dropped after the March–April peak. From July to the end of August the number of new cases remained low, mostly related to screening of subjects returning from vacation with swab testing. Cases increased again beginning in early September. Therefore, we considered the second wave to be the period from September 1 to December 31, 2020. During these periods, neither circulation of virus variants nor implementation of SARS-CoV-2 vaccination at the population level had yet started in Italy. Only a handful of Italians had been vaccinated by the end of December, but large scale administration started in January 2021 [17, 18].

Based on these data and the Italian population at January 1, 2020, and January 1, 2021 available at the ISTAT website [19], we computed wave-specific COVID-19 mortality during the first and second waves. For the former period, we used as reference population data the provincial population size as of January 1, 2020. For the second period and for the overall 2020 mortality we averaged the population size at January 1, 2020 and January 1, 2021 for each province. We retrieved the seroprevalence data for each Italian region and province determined through a national survey made available by ISTAT and the National Ministry of Health with reference to the period May 25, 2020 through July 15, 2020 [20].

Data analysis

We examined the relation between first and second wave province-specific crude mortality rates from COVID-19 by modeling the mean mortality rate using a restricted cubic spline regression [21], a model that fits a curvilinear pattern to the data as previously reported [22–24]. In particular, we fitted a restricted cubic spline model that weighted provincial data by population size, performing both a crude analysis and adjusting for potential confounders. Variables included in the model were aging index (i.e. a ratio between resident population aged ≥65 years and those aged ≤14 years), percentage commuting outside the municipality of residence on a daily basis, and percentage of dwellings occupied by only one resident (available at a provincial level from the National Institute of Statistics [19]). In the regression model, we used three knots at fixed percentiles (10th, 50th and 90th) of the first wave distribution, and computed a pointwise 95% confidence interval (CI) [21, 25]. We used ‘mkspline’, ‘regress’, and ‘xbrcsplinei’ routines of Stata software (v17.0, College Station-TX, 2021) for all analyses.

Results

Table 1 and Fig. 1 lists province-specific mortality rates in the investigated periods. National cumulative mortality (cumulative incidence of death) during the first wave averaged 58.2/100,000 persons, ranging from 6.5/100,000 persons in the Southern Italy, Basilicata region, to 165/100,000 persons in Northern Italy, Lombardy region. Corresponding figures for the second wave period were 67.6/100,000 persons for the national average, ranging from 19.4/100,000 in the Calabria region to 192/100,000 persons in the Aosta Valley region. In particular, seven provinces, five of the in Lombardy region (Bergamo, Brescia, Cremona, Lodi, and Pavia) and two in Emilia-Romagna region (Parma and Piacenza) experienced the highest mortality rates during the first wave. Second wave mortality was lower than first wave, always ≤100/100,000 persons and generally lower than the regional or national averages. The provinces with the highest first wave mortality, namely Piacenza (with the overall highest mortality in Italy with 333/100,000 persons), Cremona, Lodi and Bergamo, experienced the highest absolute decrease in mortality rates.

Table 1.

Number of SARS-CoV-2 cases, COVID-19 deaths and COVID-19 mortality rates (deaths/100,000/wave timeframe) in the 1st and 2nd waves in 2020 divided by province

Province/ Region Population Jan 1, 2020 Population Jan 1, 2021 Cases 1st wave Cases 2nd wave Seroprev. (%) Deaths 1st wave Deaths 2nd wave Mortality 1st wave Mortality 2nd wave All Deaths Overall Mortality
Aosta Valley 125,501 123,895 1195 5771 3.72 145 239 116 192 384 308
Aosta 125,501 123,895 1195 5771 3.72 145 239 116 192 384 308
Lombardy 10,103,969 9,966,992 91,813 368,273 7.35 16,633 8321 165 82.9 25,120 250
Bergamo 1,116,384 1,099,621 14,375 12,873 24.3 3137 193 281 17.4 3347 302
Brescia 1,268,455 1,247,583 15,626 25,468 7.63 2686 422 212 33.5 3117 248
Como 603,828 594,671 4093 29,531 2.00 587 794 97.2 133 1388 232
Cremona 358,347 351,698 6612 7664 19.7 1130 123 315 34.6 1261 355
Lecco 337,087 332,593 2831 10,303 6.66 481 236 143 70.5 724 216
Lodi 230,607 225,885 3570 6936 7.10 679 140 294 61.3 826 362
Mantua 411,062 403,585 3496 12,260 6.57 684 288 166 70.7 975 239
Milan 3,279,944 3,249,821 24,379 147,720 3.95 4252 3197 130 97.9 7509 230
Monza/Brianza 878,267 867,421 5772 42,090 4.52 979 895 112 103 1884 216
Pavia 546,515 534,951 5568 18,869 5.95 1241 543 227 100 1806 334
Sondrio 180,941 179,234 1584 6954 5.30 212 201 117 112 415 230
Varese 892,532 879,929 3907 47,605 1.71 565 1289 63.3 145 1868 211
Veneto 4,907,704 4,852,453 18,937 227,276 1.92 2028 4960 41.3 102 7079 145
Belluno 201,972 199,599 1191 13,369 1.88 114 330 56.4 164 445 222
Padua 939,672 929,520 3954 41,651 2.32 318 608 33.8 65.1 946 101
Rovigo 233,386 229,652 444 6932 2.39 36 204 15.4 88.1 240 104
Treviso 888,309 878,070 2673 45,715 1.89 322 783 36.2 88.7 1112 126
Venice 851,663 842,942 2682 35,612 1.68 299 863 35.1 102 1185 140
Verona 930,339 922,291 5127 44,073 2.23 586 1195 63.0 129 1789 193
Vicenza 862,363 850,379 2866 39,924 1.33 353 977 40.9 114 1362 159
Emilia-Romagna 4,467,118 4,445,549 28,061 137,052 2.90 4353 3431 97.4 77.0 7825 176
Bologna 1,017,806 1,019,539 5229 32,314 2.33 732 936 71.9 91.9 1681 165
Ferrara 344,840 341,967 1044 7886 0.72 173 222 50.2 64.6 396 115
Forlì-Cesena 394,833 393,556 1740 10,213 1.04 196 164 49.6 41.6 362 91.8
Modena 707,292 704,672 3873 25,945 1.10 480 601 67.9 85.1 1085 154
Parma 453,930 453,604 3657 8701 5.84 901 215 199 47.4 1119 247
Piacenza 287,236 284,075 4428 10,187 9.54 956 252 333 88.2 1213 425
Ravenna 389,634 386,309 1030 11,337 1.18 81 430 20.8 111 514 133
Reggio nell’Emilia 531,751 526,349 4913 18,248 4.45 581 306 109 57.8 891 168
Rimini 339,796 335,478 2147 12,221 2.79 253 305 74.5 90.3 564 167
Piedmont 4,341,375 4,273,210 30,989 162,730 3.45 4029 3537 92.8 82.1 7583 176
Alessandria 419,037 411,922 4063 13,240 2.08 659 470 157 113 1134 273
Asti 213,216 209,648 1874 7960 2.13 249 217 117 103 466 220
Biella 174,384 171,838 1046 5748 6.59 194 112 111 64.7 306 177
Cuneo 586,568 582,353 2862 24,081 0.87 373 494 63.6 84.5 869 149
Novara 368,040 362,199 2792 12,443 5.21 367 272 99.7 74.5 642 176
Turin 2,252,379 2,212,996 15,889 87,788 3.58 1844 1712 81.9 76.7 3561 160
Verbano-Cusio-Ossola 157,455 155,065 1140 5515 9.05 132 122 83.8 78.1 254 163
Vercelli 170,296 167,189 1323 5955 3.52 211 138 124 81.8 351 208
Trentino-South Tyrol 1,074,819 1,078,460 7502 43,303 9.93 693 1041 64.5 96.7 1734 161
Bolzano 532,080 533,715 2639 26,559 2.95 288 504 54.1 94.6 792 149
Trento 542,739 544,745 4863 16,744 3.42 405 537 74.6 98.8 942 173
Friuli-Venezia Giulia 1,211,357 1,198,753 3308 45,651 1.02 362 1426 29.9 118 1794 149
Gorizia 139,206 136,809 216 5904 0.12 5 104 3.6 75.4 109 79.0
Pordenone 312,619 309,058 702 9792 1.88 68 291 21.8 93.6 360 116
Trieste 233,276 229,470 1393 9107 0.59 209 270 89.6 117 483 209
Udine 526,256 523,416 997 20,848 0.93 80 761 15.2 145 842 160
Liguria 1,543,127 1,509,805 9473 46,958 3.24 1563 1276 101 86.3 2851 187
Genoa 835,829 816,916 5573 29,304 3.61 943 853 113 103 1802 218
Imperia 213,919 208,585 1494 4806 2.39 231 79 108 37.4 312 148
La Spezia 219,196 215,538 860 7142 1.89 159 189 72.5 86.9 349 161
Savona 274,183 268,766 1546 5706 3.83 230 155 83.9 57.1 388 143
Tuscany 3,722,729 3,668,333 9779 108,429 0.90 1088 2491 29.2 67.4 3604 97.5
Arezzo 341,766 336,870 676 9779 1.23 47 168 13.8 49.5 218 64.2
Florence 1,004,298 986,001 3192 29,864 0.53 401 839 39.9 84.3 1249 126
Grosseto 220,785 218,538 396 3708 1.18 28 73 12.7 33.2 102 46.4
Livorno 333,509 329,590 477 8079 0.56 62 195 18.6 58.8 260 78.4
Lucca 388,678 380,676 1351 11,010 0.42 151 192 38.8 49.9 347 90.2
Massa and Carrara 193,934 189,841 1051 6442 0 153 179 78.9 93.3 336 175
Pisa 422,310 416,425 930 15,667 1.55 91 341 21.5 81.3 433 103
Pistoia 293,059 290,819 747 9640 0.96 76 199 25.9 68.2 275 94.2
Prato 258,152 256,047 532 9790 1.02 47 203 18.2 79.0 250 97.2
Siena 266,238 263,526 427 4450 2.17 32 102 12.0 38.5 134 50.6
Umbria 880,285 865,013 1385 26,064 0.670 80 530 9.10 60.7 610 69.9
Perugia 655,403 643,311 1008 19,843 0.71 51 369 7.80 56.8 420 64.7
Terni 224,882 221,702 377 6221 0.55 29 161 12.9 72.1 190 85.1
Marches 1,518,400 1,501,406 6549 33,194 2.59 987 720 65.0 47.7 1709 113
Ancona 469,750 465,023 1875 9711 2.16 218 185 46.4 39.6 403 86.2
Ascoli Piceno 206,363 204,575 290 4790 4.95 12 125 5.8 60.8 137 66.7
Fermo 173,004 170,248 473 4337 2.16 67 69 38.7 40.2 137 79.8
Macerata 312,146 307,421 1154 7851 2.16 145 159 46.5 51.3 305 98.5
Pesaro and Urbino 357,137 354,139 2757 6505 4.95 545 182 153 51.2 727 204
Lazio 5,865,544 5,720,796 8010 148,533 1.00 863 2815 14.7 48.6 3717 64.2
Frosinone 485,241 473,467 663 12,990 0.19 79 162 16.3 33.8 241 50.3
Latina 576,655 561,139 607 13,625 0.50 44 294 7.6 51.7 340 59.8
Rieti 154,232 151,668 411 4565 3.00 41 149 26.6 97.4 191 125
Rome 4,333,274 4,227,588 5872 108,988 1.05 672 2016 15.5 47.1 2724 63.6
Viterbo 316,142 306,934 457 8365 1.52 27 194 8.5 62.3 221 70.9
Abruzzo 1,305,770 1,285,256 3261 31,124 1.29 461 794 35.3 61.3 1264 97.6
Chieti 383,189 376,397 818 6284 1.40 131 136 34.2 35.8 270 71.1
L’Aquila 296,491 292,356 225 10,604 0.54 11 350 3.7 119 362 123
Pescara 318,678 314,689 1586 5447 1.69 239 116 75.0 36.6 360 114
Teramo 307,412 301,814 632 8789 1.48 80 192 26.0 63.0 272 89.3
Molise 302,265 296,547 426 5971 0.81 28 175 9.3 58.4 203 67.8
Campobasso 218,679 214,629 364 3829 0.66 22 110 10.1 50.8 132 60.9
Isernia 83,586 81,918 62 2142 1.19 6 65 7.2 78.5 71 85.8
Campania 5,785,861 5,679,759 4648 182,462 0.89 517 2915 8.9 50.8 3447 60.1
Avellino 413,926 405,963 552 8289 0 62 143 15.0 34.9 206 50.3
Benevento 274,080 269,233 209 4423 0 19 137 6.9 50.4 156 57.4
Caserta 922,171 911,606 543 33,741 1.48 53 540 5.7 58.9 596 65.0
Naples 3,082,905 3,017,658 2652 111,294 1.04 314 1811 10.2 59.4 2133 69.9
Salerno 1,092,779 1,075,299 692 24,715 0.31 69 284 6.3 26.2 356 32.8
Apulia 4,008,296 3,926,931 4502 84,951 0.88 566 2037 14.1 51.3 2614 65.9
Bari 1,249,246 1,222,818 1491 33,237 1.50 153 636 12.2 51.5 793 64.2
Barletta-Andria-Trani 388,390 382,685 380 10,058 0.77 66 295 17.0 76.5 361 93.6
Brindisi 390,456 382,454 659 5795 0.85 67 100 17.2 25.9 167 43.2
Foggia 616,310 601,419 1170 18,639 1.02 161 655 26.1 108 819 135
Lecce 791,122 777,507 521 6420 0.01 85 114 10.7 14.5 200 25.5
Taranto 572,772 560,048 281 10,802 0.67 34 237 5.9 41.8 274 48.4
Basilicata 556,934 547,579 400 10,055 0.72 36 214 6.5 38.8 251 45.4
Potenza 360,936 354,122 189 6739 0.83 27 156 7.5 43.6 184 51.5
Matera 195,998 193,457 211 3316 0.50 9 58 4.6 29.8 67 34.4
Calabria 1,924,701 1,877,728 1179 22,191 0.51 129 368 6.7 19.4 497 26.1
Catanzaro 354,851 346,514 214 3134 0.40 31 49 8.7 14.0 80 22.8
Cosenza 700,385 684,786 468 6676 0.78 48 176 6.9 25.4 224 32.3
Crotone 170,718 166,617 119 2065 0.11 10 35 5.9 20.8 45 26.7
Reggio di Calabria 541,278 526,586 294 8586 0.18 29 81 5.4 15.2 110 20.6
Vibo Valentia 157,469 153,225 84 1730 1.12 11 27 7.0 17.4 38 24.5
Sicily 4,968,410 4,840,876 3056 89,352 0.37 342 2390 6.9 48.7 2747 56.0
Agrigento 429,611 419,847 135 3651 0.19 24 107 5.6 25.2 131 30.8
Caltanissetta 260,779 252,803 186 3733 0 18 81 6.9 31.5 99 38.6
Catania 1,104,974 1,066,765 779 26,464 0.26 103 849 9.3 78.2 957 88.1
Enna 162,368 158,183 438 2866 0 34 76 20.9 47.4 110 68.6
Messina 620,721 609,223 474 10,246 0.32 59 136 9.5 22.1 195 31.7
Palermo 1,243,328 1,214,291 500 24,929 0.89 43 665 3.5 54.1 708 57.6
Ragusa 321,215 314,950 87 6490 0.30 6 161 1.9 50.6 168 52.8
Siracusa 397,037 386,451 321 5112 0.14 47 162 11.8 41.4 218 55.6
Trapani 428,377 418,363 136 5861 0 8 153 1.9 36.1 161 38.0
Sardinia 1,630,474 1,598,225 1366 28,920 0.50 145 712 8.9 44.1 858 53.1
Cagliari 430,914 420,117 253 6573 0.38 19 161 4.4 37.8 180 42.3
Nuoro 206,843 202,951 78 6055 0.24 12 123 5.8 60.0 135 65.9
Oristano 156,078 153,226 61 2416 0.43 8 51 5.1 33.0 59 38.2
Sassari 489,634 481,052 875 8997 0.78 90 247 18.4 50.9 338 69.6
South Sardinia 347,005 340,879 99 4879 0.42 16 130 4.6 37.8 146 42.4
Italy 60,244,639 59,257,566 235,839 1,808,260 2.49 35,048 40,392 58.2 67.6 75,891 127
Open in a new tab

Fig. 1.

Fig. 1

Open in a new tab

COVID-19 mortality in the Italian provinces in the first two pandemic waves and total 2020 mortality

In the spline analysis, the relation between mortality in the first and second wave was U-shaped (Fig. 2), with a direct association between the estimates of the two waves up to 88 deaths /100,000, and above that an inverse pattern, with low second wave figures for those provinces with the highest rates during the first wave (88 through 192 deaths/100,000). The latter corresponded to a mortality rate of 22/100,000/month, as only 5 deaths were recorded in February 2020 following the first disease diagnosis in Italian residents on February 20. Crude analysis, ignoring control of potential confounders such as proportion elderly, living alone, and degree of mobility, gave similar results. When we repeated the main analysis by limiting the second wave to September–October 2020, i.e. by removing the November–December period when region-specific public health and social distancing policies were first allowed and implemented, results were not substantially changed, though mortality in this truncated period was much lower (Fig. 2).

Fig. 2.

Fig. 2

Open in a new tab

Population-weighted cubic spline regression analysis of the relation between first (from February 24 to June 30, 2020) and (A) second wave (from September 1 to December 31, 2020) mortality rates of SARS-CoV-2 infection in Italian provinces or (B) second wave before the beginning of mobility restrictions (from September 1 to October 31, 2020) mortality rates of SARS-CoV-2 infection in Italian provinces adjusted for aging index, percentage commuting outside the municipality of residence on a daily basis, and percentage of dwellings occupied by only one resident. Shadow area is the 95% confidence interval of the predicted mean mortality rate (solid line); circles are province-specific values weighted for population size

Discussion

Italy was the first country where the SARS-CoV-2 infection swept out of control, following the initial outbreak in China, where it was kept under control and geographically confined. The early arrival of the pandemic in Italy allowed a longer time frame to monitor and study the behavior of the epidemic. Italy did not allow the adoption of area-specific policies to curtail the outbreak until November, using instead a nationwide approach imposed by the Italian Government. It has been suggested recently that severity of COVID-19 pneumonia in hospitalized subjects was lower during the second wave [26], possibly due to SARS-CoV-2 genomic variation [27]. However, during the time frame of this study there was no evidence of spread of SARS-CoV-2 variants [17], thus avoiding any confounding effect from differential geographic spread of viral lineages with different transmission or virulence features. In addition, recent assessment of fatality rates between first and second waves in Italy showed that they were comparable [28]. Nonetheless, we cannot rule out that improvement in COVID-19 treatment may have decreased mortality during the pandemic, especially in the second wave, despite the lack of fully effective treatments [29–31]. Finally, vaccination had no role in curbing either of the two waves analyzed here, since almost no vaccine doses were administered in Italy in 2020. Indeed, the vaccination campaign had not begun yet in the investigated period, apart from a handful of ‘demonstration’ vaccinations administered to health professionals on December 27, 2020 at the national hub of the Spallanzani Hospital in Rome. Large scale administration started in January 2021 [32].

In contrast to our earlier analysis that relied on SARS-CoV-2 infection incidence, when area-specific mortality was still unavailable, we used mortality from COVID-19 to gauge the severity of the waves.

Mortality is a far more reliable indicator than incidence for documenting the actual trends of the pandemic, because the detection and notification of newly-infected cases from the general population depends on the availability and implementation of molecular testing, which in turn depends on health authority guidelines and policies [33]. Therefore, the number and type of individuals to be subject to SARS-CoV-2 testing may differ considerably across geographic areas or time periods independently of real differences in incidence, due to changing referral guidelines (such as the presence of COVID-19 symptoms) and availability of molecular tests. In Italy, first wave health policy allowed COVID-19 molecular testing only for COVID-19 symptomatic individuals, while in the second wave much more testing occurred, with systematic extension to asymptomatic or weakly symptomatic individuals and therefore a great increase in number of notifications. As a consequence, the ‘official’ SARS-CoV-2 infection incidence strongly increased in the second wave. This increase, however, could have been an artifact due to the aforementioned changes in testing policy and availability, especially in light of the roughly comparable numbers of deaths in the two waves. However, incidence is less reliable compared to mortality in documenting the real trend of the pandemic, also due to its inherent differences compared with SARS-CoV-2 infection incidence and with the hybrid pattern of the related assessments at the population level, due to changes in molecular testing availability and policy.

We were able to confirm the previously reported incidence patterns for SARS-CoV-2 in Italy during 2020 [12]. The key findings were a positive association between the two waves when the epidemic was not particularly strong in the first part of the year, and a negative association when the first wave was severe. This finding was similar when using a different time frame for the second wave to account for the possible influence of local policies. Local policies were permitted beginning as of November 6, 2020. In addition, the pattern we investigated is based on ‘real’ data, compared to studies presenting predictive modeling [34, 35], thus strengthening our findings.

Provinces that were relatively lightly affected by the first wave had a similar experience in the second wave, with rates closely mirroring those experienced during the first phase of the outbreak. In contrast, the Italian provinces severely hit by the first wave, such as Bergamo, Lodi, Cremona and Piacenza, experienced a marked reduction of mortality in the second wave, and presumably in the spread of COVID-19. The reasons for such an inverse pattern, however, were not directly identifiable through this study and may only be hypothesized. One possibility is a higher likelihood of immunity for superspreaders and individuals with high mobility and propensity to transmission during the first wave [36]. Another possible explanation is the development of an additional cellular immunity against SARS-CoV-2 in the most severely hit population during the first wave [37, 38]. The provinces most severely affected by the first wave showed a population-based seroprevalence in the 10–25% range, far below the estimated threshold for herd immunity of 50–70% [39]. It is possible that the prevalence of immunity, despite being below the herd immunity threshold, was high enough in provinces with a severe first wave to interact with behavioral factors and lessen the intensity of the second wave.

Acknowledgements

not applicable.

Authors’ contributions

MV and TF designed the original study, and with KJR analyzed and interpreted the data, and wrote the manuscript. NO designed and carried out data analysis with TF and SDF. All authors read and approved the final manuscript.

Funding

This study was supported by the grants ‘UNIMORE FAR 2019 and 2020 Interdisciplinare’ of the Fondazione di Modena’ to Dr. Vinceti and Dr. Filippini and by the FISR 2020-COVID19 grant by the Italian Ministry of the University and the Research to Dr. Vinceti.

Availability of data and materials

We used data made freely available from public sources. All data generated or analysed during this study are included in this published article.

Declarations

Ethics approval and consent to participate

not applicable.

Consent for publication

not applicable.

Competing interests

The authors declare that they have no competing interests.

Footnotes

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

  • 1.Kault D. Superspreaders, asymptomatics and COVID-19 elimination. Med J Aust. 2020;213(10):447–448. doi: 10.5694/mja2.50835. [DOI] [PubMed] [Google Scholar]
  • 2.Xia Y, Zhong L, Tan J, Zhang Z, Lyu J, Chen Y, Zhao A, Huang L, Long Z, Liu NN, Wang H, Li S. How to understand "herd immunity" in COVID-19 pandemic. Front Cell Dev Biol. 2020;8:547314. doi: 10.3389/fcell.2020.547314. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Okell LC, Verity R, Watson OJ, Mishra S, Walker P, Whittaker C, Katzourakis A, Donnelly CA, Riley S, Ghani AC, Gandy A, Flaxman S, Ferguson NM, Bhatt S. Have deaths from COVID-19 in Europe plateaued due to herd immunity? Lancet. 2020;395(10241):e110–e1e1. doi: 10.1016/S0140-6736(20)31357-X. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Randolph HE, Barreiro LB. Herd immunity: understanding COVID-19. Immunity. 2020;52(5):737–741. doi: 10.1016/j.immuni.2020.04.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Kwok KO, Lai F, Wei WI, Wong SYS, Tang JWT. Herd immunity - estimating the level required to halt the COVID-19 epidemics in affected countries. J Inf Secur. 2020;80(6):e32–ee3. doi: 10.1016/j.jinf.2020.03.027. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Omer SB, Yildirim I, Forman HP. Herd immunity and implications for SARS-CoV-2 control. JAMA. 2020;324(20):2095–2096. doi: 10.1001/jama.2020.20892. [DOI] [PubMed] [Google Scholar]
  • 7.Papachristodoulou E, Kakoullis L, Parperis K, Panos G. Long-term and herd immunity against SARS-CoV-2: implications from current and past knowledge. Pathog Dis. 2020;78:ftaa025. doi: 10.1093/femspd/ftaa025. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Marques M, Rovira J, Nadal M, Domingo JL. Effects of air pollution on the potential transmission and mortality of COVID-19: a preliminary case-study in Tarragona Province (Catalonia, Spain) Environ Res. 2021;192:110315. doi: 10.1016/j.envres.2020.110315. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Filippini T, Rothman KJ, Cocchio S, Narne E, Mantoan D, Saia M, Goffi A, Ferrari F, Maffeis G, Orsini N, Baldo V, Vinceti M. Associations between mortality from COVID-19 in two Italian regions and outdoor air pollution as assessed through tropospheric nitrogen dioxide. Sci Total Environ. 2021;760:143355. doi: 10.1016/j.scitotenv.2020.143355. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Cherrie M, Clemens T, Colandrea C, Feng Z, Webb DJ, Weller RB, Dibben C. Ultraviolet a radiation and COVID-19 deaths in the USA with replication studies in England and Italy. Br J Dermatol. 2021;185(2):363–370. doi: 10.1111/bjd.20093. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Filippini T, Rothman KJ, Goffi A, Ferrari F, Maffeis G, Orsini N, Vinceti M. Satellite-detected tropospheric nitrogen dioxide and spread of SARS-CoV-2 infection in northern Italy. Sci Total Environ. 2020;739:140278. doi: 10.1016/j.scitotenv.2020.140278. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Vinceti M, Filippini T, Rothman KJ, Di Federico S, Orsini N. SARS-CoV-2 infection incidence during the first and second COVID-19 waves in Italy. Environ Res. 2021;197:111097. doi: 10.1016/j.envres.2021.111097. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Romagnani P, Gnone G, Guzzi F, Negrini S, Guastalla A, Annunziato F, Romagnani S, De Palma R. The COVID-19 infection: lessons from the Italian experience. J Public Health Policy. 2020;41(3):238–244. doi: 10.1057/s41271-020-00229-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Presidente del Consiglio dei Ministri. DPCM 6 Novembre 2020. 2020 http://www.governo.it/sites/new.governo.it/files/dpcm_20200322.pdf [Accessed August 30, 2021].
  • 15.Vinceti M, Filippini T, Rothman KJ, Ferrari F, Goffi A, Maffeis G, Orsini N. Lockdown timing and efficacy in controlling COVID-19 using mobile phone tracking. EClinicalMedicine. 2020;25:100457. doi: 10.1016/j.eclinm.2020.100457. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.ISTAT. COVID-19 deaths in Italy. Report March 5, 2021. https://www.istat.it/it/files//2020/03/tabella-decessi-provinciali_fonte_istat_decessi_provinciali_per_covid_fonte_ISS_5marzo.xlsx [Accessed August 30, 2021].
  • 17.ELIXIR-IT. SARS-CoV-2 lineages of potential epidemiological interest in Italy. 2021 https://www.covid19dataportal.it/ [Accessed August 30, 2021].
  • 18.Ministry of Health. Anti-COVID vaccination campaign. 2021 www.salute.gov.it [Accessed August 30, 2021].
  • 19.ISTAT. Populations and households. Italian National Institute of Statistics. 2020 https://www.istat.it/en/population-and-households [Accessed August 31, 2021].
  • 20.ISTAT. First results of the seroprevalence investigation on SARS-COV-2. Italian National Institute of Statistics. 2020 https://www.istat.it/it/files//2020/08/ReportPrimiRisultatiIndagineSiero.pdf [Accessed August 30, 2021].
  • 21.Harrell FE. Regression modeling strategies: with applications to linear models, logistic regression, and survival analysis. New York: Springer; 2001. [Google Scholar]
  • 22.Filippini T, Adani G, Malavolti M, Garuti C, Cilloni S, Vinceti G, Zamboni G, Tondelli M, Galli C, Costa M, Chiari A, Vinceti M. Dietary habits and risk of early-onset dementia in an Italian case-control study. Nutrients. 2020;12(12):3682. doi: 10.3390/nu12123682. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Filippini T, Mandrioli J, Malagoli C, Costanzini S, Cherubini A, Maffeis G, Vinceti M. Risk of amyotrophic lateral sclerosis and exposure to particulate matter from vehicular traffic: a case-control study. Int J Environ Res Public Health. 2021;18(3):973. doi: 10.3390/ijerph18030973. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Vinceti M, Rothman KJ, Crespi CM, Sterni A, Cherubini A, Guerra L, Maffeis G, Ferretti E, Fabbi S, Teggi S, Consonni D, de Girolamo G, Meggiato A, Palazzi G, Paolucci P, Malagoli C. Leukemia risk in children exposed to benzene and PM10 from vehicular traffic: a case-control study in an Italian population. Eur J Epidemiol. 2012;27(10):781–790. doi: 10.1007/s10654-012-9727-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Orsini N. XBRCSPLINE: Stata module to tabulate differences in predicted responses after restricted cubic spline models: Boston College Department of economics. 2009. [Google Scholar]
  • 26.Borghesi A, Golemi S, Carapella N, Zigliani A, Farina D, Maroldi R. Lombardy, northern Italy: COVID-19 second wave less severe and deadly than the first? A preliminary investigation. Infect Dis (Lond) 2021;53(5):370–375. doi: 10.1080/23744235.2021.1884745. [DOI] [PubMed] [Google Scholar]
  • 27.Toyoshima Y, Nemoto K, Matsumoto S, Nakamura Y, Kiyotani K. SARS-CoV-2 genomic variations associated with mortality rate of COVID-19. J Hum Genet. 2020;65(12):1075–1082. doi: 10.1038/s10038-020-0808-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Filippini T, Zagnoli F, Bosi M, Giannone ME, Marchesi C, Vinceti M. An assessment of case-fatality and infection-fatality rates of first and second COVID-19 waves in Italy: COVID-19 fatality rate in Italy. Acta Biomed. 2021;92:e2021420. doi: 10.23750/abm.v92iS6.12241. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.COVID-19 RISK and Treatments (CORIST) Collaboration. Use of hydroxychloroquine in hospitalised COVID-19 patients is associated with reduced mortality: findings from the observational multicentre Italian CORIST study. Eur J Intern Med. 2020;82:38–47. 10.1016/j.ejim.2020.08.019. [DOI] [PMC free article] [PubMed]
  • 30.Docea AO, Tsatsakis A, Albulescu D, Cristea O, Zlatian O, Vinceti M, Moschos SA, Tsoukalas D, Goumenou M, Drakoulis N, Dumanov J, Tutelyan V, Onischenko G, Aschner M, Spandidos D, Calina D. A new threat from an old enemy: reemergence of coronavirus (review) Int J Mol Med. 2020;45:1631–1643. doi: 10.3892/ijmm.2020.4555. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Di Castelnuovo A, Costanzo S, Antinori A, Berselli N, Blandi L, Bonaccio M, Bruno R, Cauda R, Gialluisi A, Guaraldi G, et al. Lopinavir/ritonavir and darunavir/cobicistat in hospitalized COVID-19 patients: findings from the multicenter Italian CORIST study. Front Med (Lausanne) 2021;8:639970. doi: 10.3389/fmed.2021.639970. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.CPD - Italian Civil Protection Department. Anti COVID-19 Vaccine Report. 2021 https://italia.github.io/covid19-dashboard-vaccini/ [Accessed August 30, 2021].
  • 33.De Matteis A, Turkmen Ceylan FB, Adrianopoli M, Ertek S. The contribution of testing in the fight against COVID-19: evidence from Italy. Eur J Pub Health. 2021;31(4):913–917. doi: 10.1093/eurpub/ckab064. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Qian Y, Xie W, Zhao J, Xue M, Liu S, Wang L, Li W, Dai L, Cai Y. Investigating the effectiveness of re-opening policies before vaccination during a pandemic: SD modelling research based on COVID-19 in Wuhan. BMC Public Health. 2021;21(1):1638. doi: 10.1186/s12889-021-11631-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Ahmed DA, Ansari AR, Imran M, Dingle K, Bonsall MB. Mechanistic modelling of COVID-19 and the impact of lockdowns on a short-time scale. PLoS One. 2021;16(10):e0258084. doi: 10.1371/journal.pone.0258084. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Signorelli C, Zucchi A, Tersalvi CA, Ciampichini R, Beato E, Balzarini F, Odone A, Middleton J. High seroprevalence of SARS-CoV-2 in Bergamo: evidence for herd immunity or reason to be cautious? Int J Public Health. 2020;65(9):1815–1817. doi: 10.1007/s00038-020-01524-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Braun J, Loyal L, Frentsch M, Wendisch D, Georg P, Kurth F, Hippenstiel S, Dingeldey M, Kruse B, Fauchere F, Baysal E, Mangold M, Henze L, Lauster R, Mall MA, Beyer K, Röhmel J, Voigt S, Schmitz J, Miltenyi S, Demuth I, Müller MA, Hocke A, Witzenrath M, Suttorp N, Kern F, Reimer U, Wenschuh H, Drosten C, Corman VM, Giesecke-Thiel C, Sander LE, Thiel A. SARS-CoV-2-reactive T cells in healthy donors and patients with COVID-19. Nature. 2020;587(7833):270–274. doi: 10.1038/s41586-020-2598-9. [DOI] [PubMed] [Google Scholar]
  • 38.Mateus J, Grifoni A, Tarke A, Sidney J, Ramirez SI, Dan JM, Burger ZC, Rawlings SA, Smith DM, Phillips E, Mallal S, Lammers M, Rubiro P, Quiambao L, Sutherland A, Yu ED, da Silva Antunes R, Greenbaum J, Frazier A, Markmann AJ, Premkumar L, de Silva A, Peters B, Crotty S, Sette A, Weiskopf D. Selective and cross-reactive SARS-CoV-2 T cell epitopes in unexposed humans. Science. 2020;370(6512):89–94. doi: 10.1126/science.abd3871. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Fontanet A, Cauchemez S. COVID-19 herd immunity: where are we? Nat Rev Immunol. 2020;20(10):583–584. doi: 10.1038/s41577-020-00451-5. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

We used data made freely available from public sources. All data generated or analysed during this study are included in this published article.

Articles from BMC Public Health are provided here courtesy of BMC

RESOURCES