Nearly half (46%) of the important geographical differences in SARS-CoV-2 infection during the first pandemic wave in Spain may be explained by the altitude of the province of residence, and this proportion increased up to 55% when population density was also considered. The seroprevalence of SARS-CoV-2 antibodies was threefold higher in people living in provinces with altitude above the median.
It has been suggested that geographical differences in cumulative incidence of COVID-19 could be indicative of the success of preventive measures implemented [6, 7]. However, in the first wave in Spain, a decisive part of SARS-CoV-2 transmission happened before the relevant preventive measures had been introduced. Spain has coastline almost all around and many mountain ranges. Some nearby provinces have large altitude differences, while some distant provinces have similar altitude; this geographical pattern seems to have carried over to SARS-CoV-2 infection .
Other ecological studies have evaluated the relationship between environmental factors and SARS-CoV-2 infection [2, 11–14]; however, comparability among Spanish provinces was optimal in the present study because seroprevalence was assessed with the same protocol, interventions were simultaneously introduced, and social differences within the same country are probably smaller.
Studies in Latin America have not found altitude associated with higher transmission [11–14]. but the wider altitude range and tropical latitude make difficult the comparison with our results. In the present study, altitude was more predictive than any meteorological parameter, suggesting a complex synergy of several parameters on SARS-CoV-2 transmission. In Spain, an increase in altitude in March was associated with lower absolute humidity and temperature, and both factors have been related to increases in respiratory virus transmission [2–4]. Places with such conditions may have additional difficulties for effective control of transmission and may need more strict preventive measures to achieve similar results. As the present study sugests, environmental factors may modify the efficiency of SARS-CoV-2 transmission, but they are not sufficient to stop the spread; therefore, preventive interventions are necessary for all regions [5, 7, 15].
Our results of the bivariate analysis were consistent with worldwide or large countries studies that did not find association between population density and SARS-CoV-2 spread [2, 14]. However, a higher population density in the province was an independent predictor of SARS-CoV-2 seroprevalence when altitude was adjusted for. Population density has been described as a relevant factor for SARS-CoV-2 transmission in other studies [16–18]; therefore, some effect of population density on transmission in the period before the lockdown seems plausible.
The main strength of this study is that Spain met special conditions in the first pandemic wave to evaluate this association, since it is an environmentally diverse territory, the lockdown was implemented simultaneously in all regions, and a nationwide survey has been performed to obtain comparable estimates of the seroprevalence of SARS-CoV-2 by provinces.
This study has limitations. The ecological analysis may suggest a hypothesis but cannot conclude causality. Parameters of the capital city were assigned to the whole province; however, parameters of the capital city are usually a good proxy of the average conditions to which the population of the province is exposed. These results may not predict infections in further waves because a part of the population may have acquire immunity, many preventive interventions have been introduced and the weather conditions in other months may be different.