Abstract: PostScript  Open access Letter relationship between mortality and latitude below a threshold and that thereafter mortality increased with COVID-19 mortality latitude. Mortality data were log increases with northerly transformed, and piecewise linear modelling was used to explore the latitude after adjustment relationship with latitude. This was for age suggesting a link adjusted for %≥65, and pollution with ultraviolet and and population density were investigated to see if they further explained vitamin D variability in mortality. The analysis supported the hypothesis with a threshold of 28° Dear Editors, We read with interest the review north and a model of zero slope by Dr Kohlmeier in which he below the threshold, and a linear reported a correlation between model above the threshold was COVID-19 mortality among African-­ fitted. The age adjustment was Americans across the USA and highly significant (p<0.0005), with northern latitude.1 We previously an estimated mortality increase of reported a north–south gradient 13.7% (95% CI 7.4% to 20.3%) for in global COVID-19 mortality but each 1% increase in %≥65. Latitude were conscious that lack of ultra- was also significant (p=0.031) with violet exposure and consequent an estimated 4.4% (95% CI 0.4% to vitamin D insufficiency was not the 8.5%) increase in mortality for each only possible explanation.2 We have 1° further north (table 1, figure 1). now investigated the relationships Countries with higher pollution between latitude, age of population, included many with younger popupopulation density and pollution lations, and pollution was negatively associated with mortality but added with COVID-19 mortality. COVID-19 mortality per million no significant explanatory power to by country was https://www. ​ w orldometers. ​ i nfo/​ Population density expressed per coronavirus/ on 18 May 2020.3 country was not significantly associWe included all 117 countries with ated with mortality. The proportion of older people population >1 million and ≥150 COVID-19 cases. Data by country in each country impacts greatly on for population %≥65 years, popula- COVID-19 mortality, but after adjusttion density and air pollution (parti- ment for this, a strong association cles of matter <2.5 um diameter remains across the Northern hemiµg/m3) were obtained from public sphere between latitude and higher sources.4–6 Latitude was entered COVID-19 mortality. This association for each country’s capital city. The exists above 28° north not far from hypothesis was that there was no the latitude, usually stated as 35° north, beyond which populations commonly get insufficient ultraviolet B to maintain normal vitamin D blood levels throughout winter. There are exceptions, but COVID-19 mortality correlates with reported vitamin D levels across Europe,7 and in sunnier Brazil, where mortality is rising, 28% prevalence of vitamin D deficiency is reported.8 An association between vitamin D insufficiency and COVID-19 severity is supported by substantial evidence of its impact on cytokine response to pathogens.7 A direct effect of ultraviolet light on the environmental survival of severe acute respiratory syndrome coronavirus 2 is also possible but would not explain the association between mortality and ethnicity,9 whereas people with dark skin need more ultraviolet exposure for equivalent vitamin D synthesis. This analysis supports the link between latitude and..