In an interesting new paper, Andersen et al. (2021) use the Putterman-Weil historical migration index to show that life-expectancy is lower in countries where a large proportion of that country’s population emigrated from places with more sunlight (UV-R). Ethiopians in Israel, Indians in the UK and blacks in the United States, for example, tend to have Vitamin D deficiency and higher levels of mortality and morbidity from a wide variety of diseases. The effect at the global level is small but significant, about the same order of magnitude as the effect of income, inequality, and schooling.
Lots of other things are going on globally, of course, so the authors go to some lengths to control for confounding. They show, for example, that the same relationship exists within the United States. Unfortunately, they don’t have a direct measure of Vitamin D deficiency by place but they suggest based on previous research that Vitamin D deficiency is a cause of allergies so they look for whether differences in sunlight between current location and the ancestor population (DIFFUV) can explain epinephrine autoinjector prescription rates. Admittedly, there quite a few links in that causal chain but the idea is clever!
…we utilize the link between vitamin D deficiency and anaphylaxis; the latter being a serious allergic reaction (often caused by food), which is rapid in onset and may even cause death. A growing body of evidence suggests that vitamin D deficiency is an important cause of anaphylaxis (Mullins and Camargo, 2012). Laboratory evidence, for instance, suggests several mechanisms through which vitamin D affects allergic reactions in general and anaphylaxis in particular (Camargo et al., 2007). Studies also show a clear relationship between season of birth (fall and winter, the least sunny months) and food allergy prevalence (Sharief et al., 2011). A large US survey shows higher rates of food sensitization in infants born to mothers with low vitamin D intake during pregnancy (Nwaru et al., 2010). Finally, several studies document that epinephrine (a medicine used for life-threatening allergic reactions) autoinjector prescription rates vary with latitude (proxy for exposure to sunlight) in Australia, the UK, and the USA (Peroni and Boner, 2013).
Accordingly, we propose to employ epinephrine autoinjector prescription rates (EAPRs) as a crude proxy for actual vitamin D deficiency across US states. The questions we are then able to pose are the following: Does DIFFUV predict EAPR? Does EAPR correlate with life expectancy once we omit DIFFUV? Naturally, if both answers are in the affirmative then this further supports the interpretation of our main findings. Table 7 provides answers to these questions. In the first five columns we explore whether DIFFUV is a predictor of EAPR. In interpreting EAPR as a proxy for health we also control for our baseline variables: income, inequality, and human capital, as well as regional fixed effects. As can be seen upon inspection of the said columns, DIFFUV indeed correlates with EAPR in the expected way.
Hat tip: Kevin Lewis.