Second Trimester Sunlight and Asthma: Evidence from Two Independent Studies
by Nils Wernerfelt, David J. G. Slusky and Richard Zeckhauser
Asthma prevalence in the United States has been rising for the past few decades, nearly doubling from 1980-1995 and affecting close to 25 million Americans today (NHIS). What can explain this dramatic increase? The answer is crucial to understand as apart from being widespread, asthma is also extremely costly. One frequently cited study has estimated the yearly costs from asthma – including from morbidity and mortality – to be as high as $56 billion (Barnett and Nurmagambetov 2011). A growing medical literature of association and physiological studies has suggested there might be a link between asthma and maternal Vitamin D levels in utero (Weiss and Litonjua 2011). This paper aims to arbitrage some standard reduced-form econometrics to this question and test whether Americans’ asthma rates can be explained by differences in Vitamin D while in utero. Within the economics literature, our paper fits into the line of work that has focused on the fetal origins hypothesis, which has shown that in utero conditions may substantially affect later life outcomes (Barker 1992; Barker 1995, Almond and Currie 2011).
Humans synthesize Vitamin D from sunlight exposure (Holick (2004), which provides our study with both a mechanism and an identification strategy. Widespread air conditioning, electronic entertainment, and a decrease in outside works means that Americans are spending much more time inside (both for work and for leisure). Vitamin D levels have decrease substantially in response (Ginde, Liu, and Camargo 2009), a possible contributor to the asthma increase.
Our identification strategy uses within birth month-birth location sunlight variation over time. It is well known (and therefore endogenous) that some locales are sunnier than others. However, the amount of sun in a locale this year as opposed to this time in a normal year is essentially exogenous, assuming pregnant mothers do not move for this year’s sun. We employ that variation to assess the sun exposure of a pregnant mother, and correct for the norm. We then assess whether greater sun during pregnancy protects the child subsequently from asthma.
To conduct these tests, we perform two separate analyses. They use different asthma data and sunlight data, and two non-overlapping time periods. Our asthma data comes both from survey response data on prevalence (NHIS) and emergency room discharge data for children with asthma diagnoses (HCUP). The sunlight data is from weather stations (NOAA) and satellites (NSRD), matched by birth month and residence. One study is for the 1914-1987 birth cohorts, matching up NHIS data and weather stations; the other study, for 1999-2009, matches up HCUP (emergency discharges per child born in that cohort) and satellite data. The studies respectively address – prevalence and exacerbation, old and young, individual and cohort data, survey and administrative, state and county level.
The two samples produce consistent results. The difference between a second trimester of average relative sun and one of high relative sun is a decrease in asthma prevalence of 10% and in asthma emergencies by 6%. Consistent with the medical literature, we find no effect of relative changes in sunlight in the first or third trimester. These results represent a path to large potential gains, given that Vitamin D supplements are very inexpensive, and going outside is free, and that asthma imposes heavy costs.
Our findings withstand a large number of robustness checks. We use data from the ATUS and CPS to show that while relative sunlight affects time spent outdoors, it does not affect smoking which would affect asthma through another channel. Vitamin D also has the right metabolic half-life (2-12 weeks, Mawer et al. 1971; Jones 2008) for our trimester-level treatment. Alternative specifications, including a monthly one, produce the same results. Adding controls for relative changes in local pollution, temperature, or extreme temperatures all minimally affect results. Moreover, no effects are found for placebo outcomes that one wouldn’t expect to be affected by relative sunlight. Finally, adding trimester data from before conception or after birth, yields no effect except for sunlight during the second trimester.
These results indicate that there are at least moderate health benefits from sunlight exposure to pregnant women. The mechanism of Vitamin D production is related to the mechanism of suntan and sunburn, as both come from ultraviolet radiation hitting our exposed skin. Sunscreen, therefore, in addition to preventing suntan and sunburn, also reduces Vitamin D production by 97.5% (Seaing and Leung 2010). While no use of sunscreen clearly has both long- and short-term consequences, overuse may contribute to both the rise in Vitamin D deficiency and the increase in asthma prevalence. This new perspective has already influenced policy in Australia, which has one of the highest rates of skin cancer in the world. Previously, it was mandatory for school children to wear hats when outside, regardless of the time of year. In some areas, hats are now optional in the winter, precisely because of the now recognized tradeoff between skin damage and Vitamin D deficiency. We hope that our work can advance the conversation toward other benefits and costs of sun exposure.
Almond, D, and J Currie. 2011. Killing me softly: The fetal origins hypothesis. The Journal of Economic Perspectives, 25(3): 153-172.
Barker, DJ. 1992. The fetal origins of adult hyptertension. Journal of Hypertension 10(suppl 7):S39-S44.
Barker, DJ. 1995. Fetal origins of coronary heart disease. BMJ 311(6998):171-4.
Barnett, SBL., and TA. Nurmagambetov. 2011. Costs of asthma in the United States: 2002-2007. Journal of Allergy and Clinical Immunology, 127(1): 145-152.
Ginde AA, MC Liu, CA Camargo Jr. 2009. Demographic differences and trends of vitamin D insufficiency in the US population, 1988-2004. Arch Intern Med.169(6):626-32
Holick, M., and M. Garabedian. 2006. Vitamin D: photobiology, metabolism, mechanism of action, and clinical applications. Primer on the metabolic bone diseases and disorders of mineral metabolism. 6th ed. Washington, DC: American Society for Bone and Mineral Research, 106-14.
Jones G. 2008. Pharmacokinetics of vitamin D toxicity. American Journal of Clinical Nutrition, 88(suppl):582S– 6S.
Mawer, EB, GA Lumb, K Schaefer, and SW Stanbury. 1971. The Metabolism of Isotopically Labelled Vitamin D, in Man: The Influence of the State of Vitamin D Nutrition, Clinical Science 40: 39-53.
Searing, DA, and DYM Leung. 2010. Vitamin D in Atopic Dermatitis, Asthma and Allergic Diseases. Immunology and Allergy Clinics of North America 30(3): 397–409.
Weiss, ST., and AA. Litonjua. 2011. “The in utero effects of maternal vitamin D deficiency: how it results in asthma and other chronic diseases.” American Journal of Respiratory and Critical Care Medicine, 183(10): 1286-1287.