The association of UV with rates of COVID-19 transmission and deaths in Mexico: the possible mediating role of vitamin D

The first COVID-19 case in Mexico was confirmed on 26 February 2020 and by May 3 the number of registered cases had risen to 30,927. However the rate of transmission varied greatly from city to city. We used data on temperature, humidity and ultraviolet radiation (UV) from 45 cities all over the country to explore whether there was an association between these variables and rates of transmission and rates of accumulation of COVID-19 ascribed deaths. The advantage of an in-country study of this kind is that many of the variables that can confound international studies are held constant (e.g. public health policies, methods of reporting, cultural, behavioural and genetic factors). Although the official statistics undoubtedly greatly underestimate the situation in Mexico due to lack of testing, they are underestimated in all cities so this should not introduce bias across the sample. We found that temperature and humidity had no discernible association with transmission rates but that UV during the transmission period was negatively correlated with rates of transmission, suggesting a sterilizing effect. UV in the January preceding the epidemic had a slightly higher association with transmission rates than UV during the transmission period itself. We also found negative associations of UV in the transmission period and in January with rate of cumulative deaths, but at lower levels of statistical significance. We conclude that in addition to a sterilizing effect during the transmission period, UV may have a physiological effect in reducing transmission and deaths due to COVID-19, most likely through the medium of vitamin D production in the body. This follows the growing body of medical evidence that vitamin D deficiency is associated with severity of COVID-19. However, we also found a negative correlation between altitude and rates of COVID-19 transmission, distinct and independent of the UV effect, which may indicate that other physiological processes are also present. In a multiple regression, altitude and UV together accounted for 18% of the variation in transmission rates between cities.